(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Transactions of the Royal Society of Edinburgh"

11 JUL 90 



TRANSACTIONS 



OF THE 



ROYAL SOCIETY OF EDINBURGH. 

VOL. XXXV. PART II.— (Nos. 8 to 18)— FOR SESSIONS 1887-88 & 1888-89. 



CONTENTS. 



■ *. ■ c? ft 



No. VIII. Histological Observation* on the Muscular Fibre and Connective Tissue of the 
Uterus during Pregnancy and the Puerperium. By T. Arthur Hblmb, M.B. 
(With a Plate), ........ 359 

IX. On some Relations between Magnetism and Twist in Iron and Nickel. Part I. 
By Cargill G. Knott, D.Sc. (Edin.), F.R.S.E., Professor of Physics, Imperial 
University, Tokyo, Japan, . . . . . . .377 

X. On the Fossil Plants in the Raven head Collection in the Free Library and 
Museum, Liverpool. By Robert Kidston, F.R.S.E., F.G.S. (Plates I. 
and II.), ......... 391 

XI. On some Fossil Plants from Teilia Quarry, Gwaenysgor, near Prestatyn, Flint- 
shire. By Robert Kidston, F.R.S.E., F.G.S. (With Two Plates), . 419 
XII. On the Behaviour of the Hydrates and. Carbonates of the Alkali-Metals, and of 
Barium, at High Temperatures, and on the Properties of Lithia and the 
Atomic Weight of Lithium. By Professor W. Dittmar, . . . 429 

XIII. On the Determination of the Cur re, on one of the Coordinate Planes, which forms 

the Outer Limit of the Positions of the Point of Contact of an Ellipsoid of 
Revolution which always touches the Three Planes of Reference. By G. Plarr, 
Docteur es-Sciences, . . . . . . .471 

XIV. On Ostracoda collected by H. B. Brady, Esq., LL.D., F.R.S., in the South Sea 

Islands. By George Stewardson Brady, M.D., LL.D., F.R.S. (Plates 

I.-IV), 489 

XV. On Benzyl Phosphines and their Derivatives. By Professor Letts and R. F. 

Blake, Esq., Queen's College, Belfast. (With a Plate), . . . 527 

XVI. On the Anatomy, Histology, and Affinities of Phreoryctes. By Frank E. 
Beddard, M.A., Prosector to the Zoological Society of London, Lecturer on 
Biology at the Medical School of Guy's Hospital. (With a Plate), . . 629 

XVII. On the Placentation of Halicore Dugong. By Sir Wm. Turner, M.B., LL.D., 
D.C.L., F.R.SS.L. and E., Professor of Anatomy in the University of 
Edinburgh. (Plates I., II., III.), ...... 641 

XVIII. Non-Alternate ± Knots, of Orders Eight and Nine. By C. 1ST. Little, of 

Nebraska State University. (With a Plate), .... 663 



(Issued 10th March 1890). 



No. 8 issued May 5, 188). 

ii "9 m May 5, 1889. 

ii 10 ii Jwfy 1, 1889. 

u 11 n September 3, 1889. 

"12 ii September 12, 1889. 

u IS i. September 12, 1889. 

■< U ii ifarc/i 5, 2<?00. 

ii i5 ii Ocfo&e?- #, i£S9. 

.1 i6 " November 25, 1889. 

» 17 ii November 25, 1889. 

a 18 ii November 25, 1889. 



( 359 ) 



VIII. — Histological Observations on the Muscular Fibre and Connective Tissue of 
the Uterus during Pregnancy and the Puerperium. By T. Arthur Helme, 
M.B. (With a Plate.) 

(Read 9th July 1888.) 

The study of a normal process in an organ so especially prone to disease as is the 
human uterus, is beset with many difficulties ; so many foreign conditions are apt to be 
present, giving a false impression as to what this process really is. That this is the case 
to no small degree, when the character of the changes which occur in the normal involu- 
tion of the uterus after child-birth is the subject of observation, is shown by the variety 
of opinions held and stated by doubtlessly competent observers ; for while one holds that 
the muscle undergoes a fatty degeneration, another holds that no such fatty change 
occurs ; while one asserts that ODly a certain number of the fibres degenerate and 
disappear, another states that so complete is the destruction, that not one single fibre 
present in the uterus before birth survives the process ; while another goes to the extent 
of saying, that the puerperal uterus can in no way be distinguished from a uterus that 
has undergone an inflammatory process.* 

These differences of opinion are probably due to the facts that the organs studied 
did not exhibit the normal process, this being masked by the presence of other foreign 
conditions, and partly that the methods of research were not complete. For the follow- 
ing various reasons then, have I thought it advisable to observe this process of normal 
involution first in the rabbit. For owing to the fortunate rarity of maternal deaths, one 
has little chance of obtaining uteri representing the various post-partum stages (and still 
less those of pregnancy), and even when obtained, these uteri are not those of women 
undergoing a normal puerperium ; on the contrary, the normal changes in the uterus are 
materially interfered with by the presence of some important process (most frequently 
some form of puerperal fever, septic or other) which causes the patient's death. 
Further, human uteri are obtainable only after several hours have elapsed since the organ 
ceased to live, and consequently many important features can in no way be recognised, 
e.g., the proliferation of cells by the process known as karyokinesis. Again, the uteri 
are generally those of multiparous women, the great majority of whom, owing to their 
surroundings and work, are peculiarly subject to chronic uterine affections, the most 
frequent of all being a chronic condition of " subinvolution." As a result, one is apt to 
fall into many errors, and mistake the appearances due to earlier disease for those of a 
normal involution. To avoid such sources of fallacy, I have taken the uteri of rabbits 
which had been previously healthy, and were pregnant for the first time. They were 

* SCHRCEDER, SCHWANGERSCHAFT, GEBURT, and WOCHENBETT, 1867, S. 174. 

VOL. XXXV. PART 8. 3 



360 MR T. ARTHUR HELME ON HISTOLOGICAL OBSERVATIONS ON 

killed at various stages, from one to thirty-six days after parturition. The uterus was 
removed while the circulation was still going on. It was then immediately divided into 
pieces of suitable size, and plunged into one of the following fluids : — 

1. Saturated solution of picric acid. 

2. Solution of chromic acid, — £ per cent. 

3. Solution of osmic acid, — \ to 1 per cent. 

4. Midler's fluid. 

5. Absolute alcohol. 

Similarly, uteri of rabbits at various stages of pregnancy, the uterus of a virgin 
rabbit, and that of a rabbit which had not borne for several months, were taken and 
treated in a like manner. 

I have divided this paper into two parts — in the first, giving a resume of the general 
literature showing our present position in regard to the changes in the human uterus and 
that of lower animals ; and in the second, dealing with the rabbit and recording the facts 
observed and conclusions come to by myself. I first give the general literature including 
that of the human uterus, for a twofold object — first, to make clearer the points to which 
I have directed my attention in the rabbit's uterus, and to contrast the process that 
occurs in the rabbit with that said to occur in the human subject ; and secondly, to 
anticipate my own work, as yet incomplete on the same subject. 



PART I. 

Literature of the Changes in the Human Uterus and that of Animals 
during Pregnancy and the Puerperium. 

In reviewing the literature, it is convenient to consider separately the human 
uterus and that of animals, and under each to take up (l) pregnancy, (2) the puerperium. 

I. The Human Uterus. 
A. Pregnancy. 

1. Muscular Tissue. — The changes that take place in the uterine musculature and 
connective tissue during pregnancy have, considering the importance of the subject, 
received very meagre attention. Few observations are recorded, and these many years 
ago. 

Kolliker asserts that the increase in size of the uterus is due to an increase of all 
its constituent elements, the process in the musculature being two-fold, namely, 

a. A new formation of muscle cells. 

b. A hypertrophy of existing cells. 



THE MUSCULAR FIBRE AND CONNECTIVE TISSUE OF THE UTERUS. 361 

During the early months, these two processes go on side by side ; after the fifth 
month the second process only is to be found.* 

These conclusions were based upon the examination of two uteri, one at the fifth 
month of pregnancy, and one at the second half of the sixth month, the muscle fibres 
of which are carefully described. 

The new muscle fibres are developed from small formative cells.t 

Eobin, so far as I have read, is the only observer who states the case differently. 
He describes only a single formative process, namely, the hypertrophy of pre-existing 
cells.| I have been unable to obtain his original articles, which I believe appeared in 
the Arch. gen. de Med.., 1848, 1858, and 1861. 

Most of the text-books (Playfair, Lusk, Schrceder, Quain, Rutherford, &c.) 
teach Kolliker's views. 

Schafer, in Quain's Anatomy, describes the increase by hypertrophy of existing 
cells, cautiously adding, — " It is said also by development of new muscular fibres."§ 

As to the origin of these new fibres most authors are silent, others adopt Kolliker's 
view, while Rutherford points out that their " origin has yet to be made out."|| 

2. Connective Tissue. — As to the connective tissue, little has been written ; all agree 
that it, along with the vessels, undergoes a great increase corresponding to and simul- 
taneous with that of the musculature. 1F 

Lately, Meola has stated that, towards the end of pregnancy, he finds a marked 
cellular development in the connective tissue, the cells remaining during pregnancy in 
embryonal form. ## 

B. Puerperium. 

1. Muscular Tissue. — The earliest allusion to the state of the muscle fibres of the 
uterus post partum that I can find, is one by Kolliker (in 1849), who describes their 
rapid diminution in length, along with the presence of fat granules. He does not give 
an account of the puerperal changes, but merely describes one uterus three days after 
labour, tt 

The next observations were those of Heschl (in 1852), who described a total 
destruction and reconstruction of the entire uterus. He based his conclusions on the 
examination of a uterus, the description of which proves undoubtedly that it was a 
pathological specimen, as he states that " the peritoneal investment is covered, two days 

* Kolliker, Mikroskopische Anatomie, 1854, Zweiter Band, Zweite halfte, s. 448-9 ; also in 4th edition, 1863, 
pp. 567-8. 

t Ze.it. f. Wissensch. Zool., 1849, Erster Band, s. 72. 

X Diet. Encycl. de. Med., 1876, 2 me sene, vol. x. p. 537. 

§ Quain's Anatomy, 1882, 9th ed., vol. ii. p. 135. 

|| Rutherford, Text-Book of Physiology, vol. i. p. 135. 

IT Kolliker, Anatomie, s. 449 (Connective Tissue) ; s. 455 (Veins and Arteries). 
** Gent.f. Gynec, No. 1, 1885, Jan. 3, from 11 Morgagni, 1884. 
tt " Beitrage zur Kentniss der glatten Muskeln, von. A. Kolliker," Zeit. f. Wissen. Zool., Erster Band, 1849, s. 73. 



362 MR T. ARTHUR HELME ON HISTOLOGICAL OBSERVATIONS ON 

after birth, with small exudations from which, later, result the adhesions of the uterus 
to the neighbouring organs."* 

Kolliker, in his Mikros. Anatomie (1854), states that his observations on the human 
uterus agree with Kilian's observations on the Mammalia during pregnancy, but differ 
on the post-partum changes. Kolliker here states that he finds fatty metamorphosis, 
but that there is not a total destruction of the uterus ; on the contrary, the great majority 
of fibres are not destroyed, but last through more than one pregnancy."t 

Robin holds that there is no destruction of fibres, merely diminution, and return to 
their original state ; he further denies the fatty metamorphosis, and describes the process 
as a simple atrophy .J 

Luschka, in his Anatomie, speaks of the muscle cells being " restored to their original 
size." This involution process being accompanied by an appearance of fat in the cells. 
He gives no ground for his opinion, merely states it without giving any description of 
material examined on which he bases this statement, which differs from all preceding. § 

After this there seems to have been a lull in the observations on this subject, for it 
is not till 1885 that one finds another paper. Meola advances a new theory, 
comparing the process to a cirrhosis ; rapid growth of connective tissue in the early 
days of the puerperium compressing the muscle cells, which diminish by a process of 
granular atrophy. I have been unable to see Meola's original article, of which an 
abstract appeared in the Cent. f. Gyn.\\ 

Lately, since I began to work at the subject, two papers have appeared, one in Germany 
and one in France. The former by Sanger of Leipsic, who describes the development 
of fat within the muscle cells (" No fat debris is found outside the muscle cells "), but 
there is no destruction of fibres, merely a return to normal. IT The last paper is by Mayor, 
in December 1887, who gives a full account of the specimens examined, with the cause 
of death of the patient. He describes the appearance of very fine fat granules — so fine 
are they in the sketches he has published as to be scarcely recognisable — but denies 
the destruction of any fibres. ## 

Lastly, I may mention a paper by Beneke, in which he describes the finding of a 
hyaline degeneration of fibres in the puerperal uterus. As a result of his experiments 
with smooth muscle of various organs of animals, he concludes that this hyaline degenera- 
tion is an artificial post-mortem production. tt 

2. Connective Tissue. — The general opinion seems to be that the connective tissue is 
reduced in a degree corresponding with that of the musculature, but the process is not 

* " Untersuchungeii uber das Verhalten des menschlichen Uterus nach der Geburt," von Dr Hesohl, Zeit. der Kais. 
him. Gesellschaft. derAerzte zu Wien, Achter Jahrgang, 1852, Bd. ii. s. 228, &c. 

+ Kolliker, Mikroskopische Anatomie, s. 452. 

X Diet. Encyc. de Med., Art. " Musculaire," 1876, 2™e sdrie, vol. x. pp. 537, 541. 

§ Luschka, Die Anatomie des Menschen, 1864, 2 Bd. 2 Abtheilung, s. 365. 

|| Cent. f. Gyn., No. 1, 1885, Jan. 3 ; and 77 Morgagni, 1884. 

IT Beitrage zur Path. Anat. und Klin. Med., Leipzig, 1888. 

** Arch, de Physiologie Normale et Pathol., No. 8, 15th Nov. 1887, p. 568, &c, Troisieme s<5rie, tome dixieme, 1887. 

tt Beneke, "Zur Lehre von der hyalinen Degen. der glatten Muskelfasern," Virch. Arch., Bd. 7(7. Heft 1, s. 71. 



THE MUSCULAR FIBRE AND CONNECTIVE TISSUE OF THE UTERUS. 363 

described. The three most important and definite statements are those by Sanger, 
Meola, and Mayor, each being directly opposed to the other. Sanger holds that the 
connective tissue undergoes an involution gradually diminishing without the appearance 
of fat, and playing an entirely passive part. 

Meola, on the other hand, ascribes a very active and important role to the connective 
tissue, which, instead of degenerating, advances from an embryonic state to one of 
complete development, the cells becoming fully formed fibrous tissue. 

Mayor, again, holds that the connective tissue cells become loaded with fat, which 
afterwards disappears, the cells acting as reservoirs, taking up the fat, and passing it on 
gradually into the circulation, from which again, he suggests, it is . stored up in the 
connective tissue cells of the body, these being transformed into fat cells. 

With regard to the post-partum changes in the vessels, a large amount has been 
written on those of the placental site, while little attention has been paid to the state of 
the vessels of the general uterine wall. Balin, however, has described a process of 
proliferation of the endothelium of the intima occurring in the vessels (other than those 
of the placental site).* 

II. Uterus of the Rabbit. 

Very little has been written on this subject. I can find only two works ; the first is 
that of Kilian, who published his researches in 1849. He describes the increase of the 
musculature during pregnancy as due to a two-fold process — namely, hypertrophy of pre- 
existing cells, and development of new cells from flat rounded cells with round nuclei.t 

I should not like to pass on to the description of my own work without drawing 
attention to one or two points which occur in the passages above referred to. 

In the first place, Kilian's methods appear to have been very primitive ; indeed, the 
onty methods he used were the following : — First, to tease up the fresh specimen, then to 
add acetic acid ; if during the first examination he found granules which did not disappear 
during the second, he concluded these granules were of a fatty nature. His other method 
was first to dry the preparation, then soften it again with water, and finally to add acetic 
acid. Now with regard to this method of teasing the fresh preparation, unless it be done 
very carefully, so as absolutely to isolate each muscle cell, one does find that the muscle 
cells appear granular, but this is because the exceedingly granular intercellular substance 
adheres to the outside of the fibres. Further, during the process of teasing, one must 
necessarily liberate fat granules present in any other tissue, and these are normally 
present, as is well known, in peritoneal cells ; and I have also found them, as described, 
in the subperitoneal connective tissue cells, so that it is quite possible for fat granules 
liberated from other tissues to float into and adhere to the muscle cells. This method 

* Balin, Archiv f. Gyn., xv. ; also Friedlander, Physiologisch-anatomische Untersuchungen uber den Uterus, 1870 ; 
Leopold, Arch. f. Gyn., xii.; Patenko, Arch.f. Gyn., xiv. s 422. 

t "Die Structur des Uterus bei Thieren," von Dr F. M. Kilian, Zeitschrift fitr rationelle Medicin, Band ix. Heft 1, 
1846 ; s. 9, 15, 32, 33-35, 40-41. 



364 MR T. ARTHUR HELME ON HISTOLOGICAL OBSERVATIONS ON 

then is insufficient to prove the existence of fat granules in the muscle cells. We require 
to have the tissues fixed, and also the assistance of osmic acid. 

Kilian assumes that, during pregnancy, new muscle cells are developed on the solitary 
basis, that he found fibres of varying size ; but during pregnancy the connective tissue 
is so greatly developed, that its cells are apt to be taken for young muscle cells. The 
nuclear movements known as karyokinesis were unknown in his day. 

With regard to the theory of involution, which he caricatures, it seems at first sight 
to resemble the description of the process which I shall have to offer, but I must point 
out that this theory, out of which Kilian derives so much amusement, is a purely 
mechanical one, the fibres being compared to sponges, as it were, which suck up the 
juices during pregnancy, afterwards squeezing them out again. This looks like a 
veritable " man of straw " specially built by Kilian for the purpose of having something 
to destroy. He bases this theory of his on " the most superficial investigation ; " the 
description I shall later give is, on the contrary, based on very careful and even prolonged 
investigation, not on the examination of two uteri, but of a dozen or more. 

A second research has been lately carried out by Bernstein, who embodied his 
results in his inaugural dissertation at Dorpat. This thesis I have not seen, being able 
only to find a short reference in Schmidt's Jahrbuch for 1886, and certain quotations in 
the papers of Sanger and Mayor. 

Bernstein investigated the puerperal involution of the uterus in rabbits, by making 
measurements of the uterine connective tissue. He comes to the conclusion that the 
connective tissue in the puerperium is absolutely diminished, but that it atrophies to 
a slighter extent than the musculature, and that a not unimportant number of con- 
nective tissue cells do not undergo fatty degeneration.* 

PART II. 

So much for the literature of the changes in the uterus. I shall now pass on to the 
second part of this paper, namely, my own observations on the rabbit's uterus, taking 
up — T. Pregnancy, II. Puerperium. 

Pregnancy. 

Here we consider the changes in the muscular and connective tissues and in the 
blood-vessels. 

1. Muscular Tissue. 

Following closely as a corollary to Virchow's dictum " Omnis cellula e cellula," it 
is believed that the cells of any one tissue are developed only from pre-existing cells 
of the same tissue — that is to say, cells have an isogenous as contradistinguished from an 
allogenous origin. With regard to non-striped muscle in particular, that this is the case 
has been beautifully shown by the observations of Stilling and Pfitzner of Strassburg. 

* Schmidt's Jahrbuch, 1886. 



THE MUSCULAR FIBRE AND CONNECTIVE TISSUE OF THE UTERUS. 365 

These observers excised small bits of the muscular coat of the Triton's stomach, and after 
allowing the process of healing to go on for various periods (a few weeks to six months), 
examined the site of injury. They found at the edges of the cut muscular layers 
numbers of muscle cells, whose nuclei presented the characteristic karyokinetic figures, 
showing that the renovation of non-striped muscle takes place by the development of 
new cells from parent muscle cells. We may then dismiss from our minds the possibility 
of any development of muscle fibres from leucocytes or embryonal elements, &c, and 
take as proof of the occurrence or non-occurrence of the development of new muscle 
fibres the presence or absence of dividing nuclei of the existing cells. Accordingly, 
preparations of the uterus at all stages of pregnancy were made with this single object 
in view — to examine the nuclei for division. The method I employed was that employed 
by Stilling and Pfitzner, and also the picric acid method. That the methods were 
not faulty is shown by the numbers of karyokinetic figures observed in the somewhat 
large nuclei of the clear glandular epithelial cells ; but in the muscular tissue not a 
single dividing nucleus was seen at any stage of pregnancy. The cells presented a 
gradually advancing growth in all diameters, in the nucleus as well as in the cell-body. 
It is also noteworthy, that as the cells increase in size they begin to present longitudinal 
striations and transverse rings — the former so marked towards the end of pregnancy 
that the transverse sections of the muscle cells may be easily mistaken for the prickle 
cells of the skin. During the last few days a very rapid increase in the size of the 
individual cells, and a marked change in their appearance, are to be noticed. The cells 
increase rapidly in volume, and assume a clearer, glassy aspect, looking as if the pro- 
toplasm were saturated and distended with water. At the same time, some of the cells 
present the appearance of " vacuoles," clearer spaces with a definite outline, generally 
by the side of the nucleus ; the latter being pushed outwards and curved, as it were, by 
this vacuole, so that it, the nucleus, seems to form part of the periphery of the vacuole. 
The appearances remind one of those of cells which have undergone what pathologists 
describe as serous degeneration — their refractile character also suggesting an early stage 
of hyaline degeneration.* 

The importance of these changes I shall speak of shortly, but with regard to the 
question at present before us, my observations show that the increase of the uterine 
musculature during pregnancy is brought about, not by a double process, but only by the 
increase in volume of the existing muscle cells. 

One uterus which T examined at the 21st day of pregnancy had only one horn 
impregnated — the unimpregnated horn, however, had undergone a similar hypertrophy 
in all its parts. 

* I find that this appearance has heen noted by Dr Julius Elischer, who, in a paper on " The Finer Anatomy 
of the Muscle-fibre of the Uterus" (Arch. f. Gynec. Band ix., 1876, s. 15), says — "My observations of the contractile 
substance show a decided difference according as the muscle of a non-pregnant, or, on the other hand, of a pregnant 
uterus is taken. Without further treatment with reagents, only teased in amniotic fluid, the muscle fibres of the non- 
pregnant uterus show the contractile substance more markedly striated, ami have an appearance (Glanz) much darker 
than those of the pregnant organ, whose appearance, I might say, reminds one of that of slight amyloid degeneration." 



366 MR T. ARTHUR HELME ON HISTOLOGICAL OBSERVATIONS ON 

2. Connective Tissue. 
In the virgin uterus the muscle bundles lie closely together, scarcely separated by any 
connective tissue, which is present only in very small amount, and is of the fully formed 
white fibrous variety, with its different form of corpuscles, and with an admixture of elastic 
fibres. When pregnancy occurs the connective tissue and vessels participate in the 
general activity, so that there is along with the enlargement of the muscle cells a 
corresponding and contemporary increase in the amount of connective tissue. But 
during the first half of pregnancy this is moderate in amount, and the process goes on 
to complete development, the cells becoming transformed into fully-developed fibrous 
tissue, so that on examining a uterus at the 

14-th day of Pregnancy, one finds the connective tissne still small in amount, but round the 
vessels there is young fibrous tissue forming, though there is no excess of cellular elements. 

16th day of Pregnancy. — The connective tissue has developed, especially in the zone between the 
two muscle layers. Young vessels too are being formed by the connective tissue cells arranging 
themselves in rows, chiefly at right angles to the circularly arranged layers of muscle cells. The 
layers of muscle fibres, however, are not much broken in upon, the muscle bundles still lying closely 
together. 

From this time the cellular character of the connective tissue begins to be more 

marked, there is a striking emigration of leucocytes from the vessels into the perivascular 

spaces; so that the cells increase largely in number, and do not go to complete formation 

of fibrous tissue. 

21st day of Pregnancy. — There has been great activity in the connective tissue. The small 
vessels are surrounded by patches of connective tissue rich in cells, small leucocytes, and larger cells 
with clear nuclei, and more protoplasm surrounding the nucleus. The connective tissue has also 
developed among the muscle layers, so that the bundles lie apart, separated by strands of connective 
tissue rich in cells. Many of these cells are transforming themselves into fibres, while others 
retain their cellular character, the nucleus enlarging and becoming clearer. 

There is also a considerable emigration of leucocytes. At this time a very active 
increase goes on in the connective tissue, especially between the two muscular coats 
(longitudinal and circular) — the new tissue and vessels push their way in among the 
muscle bundles, so that instead of the simple arrangement of one external longitudinal 
and one internal circular coat, an intermediate layer, made up of bundles interlacing 
in every direction, becomes developed. During the last three or four days there is an 
entirely new phenomenon, a new variety of cell makes its appearance — colossal cells — 
many of which are multinucleated. 

26th day of Pregnancy. — Great increase in connective tissue, with preponderance of cell 
element. A change in the character of the cells is to be noled. In addition to the ordinary 
connective tissue corpuscles and leucocytes, lying around the blood-vessels (among the connective 
tissue), are larger cells with finely granular substance, and a single clear large nucleus, which 
contains one or two very definite nucleoli. These epithelioid corpuscles, which are probably 
ordinary wandering conneotive tissue cells enlarged, form islands round the vessels, and between the 
muscle bundles. In other places new cells have appeared, large, granular, with two or three large 
clear nuclei. 

27lh. day of Pregnancy. — Great increase of the large multinuclear cells. They are four or five 



THE MUSCULAR FIBRE AND CONNECTIVE TUSSUE OF THE UTERUS. 367 

times the size of the large epithelioid connective tissue cells, but their size varies considerably ; the 
shape also varies, but they are generally spheroidal, with an even or irregular outline. The cell 
substance is very granular, but the nuclei, which vary from one to three or four in number, are 
particularly clear, and possess two or three very definite nucleoli. These huge cells are found 
chiefly in groups in the strands of connective tissue separating the muscle bundles; some of the 
cells, however, lie almost isolated, and others again seem to have wandered in among the smaller 
muscle bundles. Where these cells are in greatest number, there is a corresponding diminution 
in the number of the smaller single nucleated cells, giving the impression that these plasmodia 
are formed by growth, and then fusion of the above single nucleated cells. 

Being developed then in these large numbers at this special period, these plasmodia 
must have some definite object, and it is natural to conclude that it must be in prepara- 
tion for the puerperal changes. As to their origin, it is probable they arise from the 
smaller cells with single large clear nuclei, which again have had their origin in the 
emigrated leucocytes. 

3. Vessels. 

With regard to the development of new vessels, it is impossible to make out com- 
pletely all the processes. But that they increase by multiplication of the existing endo- 
thelial cells is demonstrated in several of my specimens (fig. 1), this multiplication 
being originated by the usual nuclear movements— not a process of budding with subsequent 
subdivision of the nucleus. It is an interstitial increase. But these dividing cells are not 
sufficiently numerous to account for the great increase of capillary vessels, and several 
preparations show appearances that suggest the new vessels are largely formed by the 
arrangement and junction of the endothelioid cells of the connective tissues. An active 
process of growth goes on in arteries, veins, and capillaries. 

An interesting fact is the sinus-like character of the veins which in many cases 
appear as periarterial sinuses (fig. 4), large spaces lying alongside of and sometimes almost 
encircling the arterial tube, and having merely a thin wall made up of endothelial cells 
and devoid of muscular tissue. 

An important point is the fact that during the last two or three days of pregnancy, 
the lining cells of the vessels and the muscle cells of the arteries undergo a rapid swelling, 
becoming clear and hyaline. This is of a most pronounced character, and is important, as 
it must tend to diminish the amount of blood supply to the contractile and connective 
tissue elements of the uterus. 

Looking then at the uterus as a whole, we find a gradual increase in its volume, due to 
the growth of all its constituent elements : the connective tissue, vessels and glands by 
an increase in their amount, i.e., a development of new elements or new formation : the 
muscular tissue by a process of simple hypertrophy of pre-existing cells. Towards the 
last days of pregnancy, we find evidences of even greater activity of a remarkable three- 
fold character : — 

1. The swelling of the muscle cells till they become distended and hyaline. 

2. The swelling of the lining endothelium and of the muscle cells of the vessels. 

3. The appearance of numerous plasmodia. 

VOL. XXXV. PART 8. 3 P 



368 MR T. ARTHUR HELME ON HISTOLOGICAL OBSERVATIONS ON 

Now these facts must have a relation to the rapidly approaching completion of the 
uterine function, i.e., parturition. 

It seems as if each cell of the uterus had its life-history distinctly laid down for it, 
and involved within itself. In the virgin state the fibres are in an undeveloped state ; 
their power of development is latent, to be called out on the application of some stimulus, 
e.g. , conception. When this occurs, each fibre develops to its utmost, and during its evolution 
brings about the reason for its own involution. The augmentation of the volume of each 
individual fibre is seen to occur with greater rapidity the nearer one gets to the end of 
pregnancy, until at the full term the enlargement is so rapid that the supporting frame- 
work which conveys the nutriment has no time to accommodate itself to the rapidly 
increasing pressure of the distending muscle cells. Consequently, the nourishment begins 
to be cut off, and at the same time a second and a third factor are at work in the same 
direction, viz., the diminution of the lumen of the vessels by a similar process occurring in 
their muscle coat and lining endothelium, and the rapid development of plasmodia 
(assisting in compressing the smaller vessels). 

As a result of their interference with its blood supply, the muscle cell becomes more 
and more irritable, so that the contractions of the uterus, which are normally occurring, 
become increased in intensity. This is well known to occur in the human female, the 
increased frequency and intensity of the uterine contractions being looked upon as one 
of the signs of approaching labour. One of the most powerful stimuli for induction of 
muscular contractions of the uterus (as I have found during the course of experiments 
with the artificial circulation of the uteri of sheep) is a momentary stoppage of the 
blood-stream. While the blood-stream is flowing, the normal contractions of the uterus 
go on rhythmically, but almost coincidently with the stoppage of the stream a violent 
contraction of much greater intensity and much increased duration occurs. Now it 
seems to me that as the result of the interference with the blood-stream, as I have 
described in the pregnant uterus, the muscle cells become more and more irritable and 
respond more and more intensely to any stimuli. The result of this increased irritability 
of the uterine muscle is the intensification of the uterine contractions, which find their 
culminating point in the act of labour. 

But in the evolution of the uterus, not only have the needs of labour been provided 
for, but also those of the puerperium. 

And here I should wish to draw attention to the sinus-like character of the thin- 
walled veins, in relation to the risk of haemorrhage. This is perhaps more important 
in the human being, where the risk is greater. In this sinus-like character of the veins 
we see a beautiful natural means for the arrest or prevention of haemorrhage. As the 
foetus is expelled, and the placenta detached, the uterine musculature contracts, and at the 
same time retracts, so that the extent of the uterine surface remains permanently smaller. 
The result of this retraction is a readjustment of or alteration in the planes of the con- 
nective tissue, and a consequent compression and blocking of the uterine sinus-like veins. 
If I may allude to the human uterus again, this is no doubt the explanation of the arrest 



THE MUSCULAR FIBRE AND CONNECTIVE TISSUE OF THE UTERUS. 369 

of bleeding in the cases of accidental haemorrhage, which immediately follows the artificial 
rupture of the membranes and partial emptying of the uterine cavity, namely, by the con- 
traction and retraction of the uterine muscle, an alteration in the connective tissue planes 
occurs, resulting in compression of the thin- walled vessels. In this too we see Nature's 
method of preventing haemorrhage, when the vessels at the surface of the uterus are torn 
through in the third stage, by the separation of placenta and decidua. 

Further, in the great development of the army of plasmodia, we see an efficient pro- 
vision for the changes of the involution process. This will be discussed later, but here 
let me emphasise the fact, that at the moment of labour there are lying in the uterine 
wall numbers of large plasmodia, whose function no doubt is called out during the 
puerperium. 

Poerperium. 

In describing the changes in the uterus during the puerperium, I shall give first a 
detailed description of the preparations which I have examined, and then the conclusions 
founded on them. 

Description of Uteri of the Rabbit at all Stages of the Puerperium. 

The rabbits were killed at various stages, from one to thirty-six days after parturition. 
The uterus was removed while the circulation was going on, and treated by various 
reagents as before described. 

I. 24 Hours Post Paetum. 

The muscle cells were examined both fresh and after hardening. 

(a) Fresh, removed from the living animal. At first, on examining the teased preparation, 
I got an appearance in which all the muscle fibres and connective tissue seemed to contain and 
be covered by small " granules " or " globules." On these potash solution (32 per cent.) had little 
or no effect. Acetic acid almost entirely cleared them off. But on making more careful pre- 
parations, and with great trouble separating the individual fibres, it was evident that these fine 
granules were not in the muscle cells, but in the surrounding connecting substance which could be 
seen adhering as an irregular fringe to the sides of some fibres — other fibres, however, being quite 
free from it. 

(b) Hardened in Miiller, stained alum carmine. The muscle cells present a cloudy dim 
appearance, so that the nuclei are indistinctly seen — the cells, in fact, are in a condition resembling 
early hyaline degeneration, being dim and hazy, but at the same time somewhat refractile, so that 
the individual fibres stand out from one another. 

(c) Hardened in osmic acid (1 per cent.). No fat granules are seen in the muscle cells. 

The connective tissue is very granular — granules being present in both the intercellular tissue 
and the cells. Some of the cells lying under the peritoneal covering contain fine fat particles. 

The large plasmodia are not seen in groups, but scattered, and instead of the rounded shape of 
pregnancy, show every variety of irregular form, some more or less rounded, others pear-shaped, 
others with processes running out in several directions. Their protoplasm contains numerous large 
and fine granules, probably albuminoid, as though their contour becomes more defined by the osmic 
acid, they are not blackened. These cells are especially numerous towards that point of the uterine 
wall where the vessels enter and emerge, are rarer and fewer between the muscle bundles ; — also 
lying around the blocked capillaries of the submucous tissue are numbers of granular cells with large 
single nucleus. 



370 MR T. ARTHUR HELME ON HISTOLOGICAL OBSERVATIONS ON 

II. 36 Hours Post Partum. 

The muscle cells were examined fresh and after hardening. 

(a) Fresh. The individual fibres show very slight, if any further diminution in size, but their 
outlines and the outlines of their nuclei are less regular. The cell substance is dimmer than during 
pregnancy, so that the cells have a more hazy look. 

(b) Osmic acid. None of the cells show any appearance of fat. The outlines of the cells and 
the cell contents are darkened to a greyish colour, but no blackened fat drops are present. 

(c) Picric or chromic acid. Gland epithelium is proliferating — typical nuclear figures in 
numbers. 

The Connective Tissue. — The intercellular substance and the ground substance of the connective 
tissue is very granular, but clears up very considerably under acetic acid. Treated with osmic 
acid, in the cells beneath the peritoneum fine fat particles are seen ; they are also very numerous 
in the peritoneal cells. The plasmodia are present in numbers, lying in the granular connective 
tissue among the muscle bundles. 

III. 2\ Days Post Partum. 

The muscle cells were examined fresh and after hardening. 

(a) Fresh. The isolated fibres are translucent, somewhat refractile ; outlines indefinite when 
several are together. Nucleus is recognised as a slightly dimmer part of the cell, with one 
(generally) or two or sometimes three bright glistening nucleoli. 

The cell substance is fairly clear, very finely granular, almost homogeneous in appearance 
(few distinct granules being outlinable). The isolated nuclei are quite homogeneous except for 
one or two shining points. 

The intercellular substance and connective tissue ground substance are very granular. 

After acetic acid, the muscle cells swell up, becoming clearer and absolutely homogeneous ; at 
first the nuclei are rendered more distinct, but later they too become pale and homogeneous, so that 
their outline can no longer be recognised. The connective tissue also swells up and becomes clearer, 
in many places not a granule remaining visible, the whole tissue being quite homogeneous in 
appearance ; in other places, however, some larger granules (few in number) still persist. 

After potash, the nucleus becomes if anything more distinct, and the cell substance a little 
clearer — perhaps only in contrast to the intercellular substance and the connective tissue, which 
takes on a more granular appearance. 

(b) Hardened in Miiller's fluid. The muscle cells all show a diminution in volume, the nuclei 
also are correspondingly smaller. The cell substance is not so refractile as in the 24 and 36 hours' 
specimens — it has lost the peculiar dim hazy look, and instead is fairly clear, but slightly granular. 

(c) Osmic acid shows presence of no fat. 

IV. 3£ Days Post Partum. 

The muscle cells, examined when fresh, are very finely granular, having the appearance of 
" roughened " glass, so that the nucleus is slightly hidden ; it becomes visible on the action of 
acetic acid or glycerine, and shows a normal appearance. 

With a magnification of 1200 diameters, fine granules are seen in the cell substance. The 
connective tissue shows a very granular appearance as before. Osmic acid shows no fat droplets in 
the muscle cells. 

The connective tissue is even more granular than before, and in the subperitoneal cells fine fat 
granules are seen. The peritoneal cells also contain a large amount of fat. 

The large granular connective tissue corpuscles are very numerous. The plasmodia are much 
fewer in number, though still to be found both between the muscle bundles and in the submucous coat. 

The gland cells are undergoing proliferation. The whole uterus is much reduced in size ; the indi- 
vidual muscle cells are still smaller than in the 2£ days' uterus, and the nuclei also are diminished. 



THE MUSCULAR FIBRE AND CONNECTIVE TISSUE OF THE UTERUS. 371 

V. 6 Days Post Partum. 

The muscle cells were examined fresh and after hardening. 

(a) Fresh. The muscle fibres are more granular (though still very finely). The nucleus is 
not so visible ; individual fibres are less easily distinguished from one another. After being in salt 
solution (0'75 per cent.) for some time (half an hour), the cells become more transparent, though still 
showing fine granules, and the nuclei become clearly visible. 

Acetic acid and potash act as before. The connective tissue ground substance is extremely 
granular. 

(b) Osmic acid shows presence of no fat. 

(c) Hardened in picric or chromic acid. The muscle cells are smaller than in the 3 days' uterus ; 
the connective tissue also is less in amount, and the whole uterus is considerably diminished. All 
the muscle cells are in an exactly similar state, no appearance suggesting that some fibres disappear, 
while others remain. 

No karyokinetic figures seen in the muscles. 

The connective tissue is much smaller in amount, being reduced almost to the amount in a 
normal non-pregnant uterus, and the fat granules have almost entirely disappeared. 

In the submucous connective tissue large granular cells with large yellow granules in their 
interior are seen. Some contain whole red blood-corpuscles. 

VI. 10 Days Post Partum. 

(a, Fresh. The muscle cells are very much smaller, the nucleus being surrounded by only a 
small amount of dim cell substance. 

The individual fibres are dimmer and more refractile than are the fibres of early pregnancy ; they 
all resemble each other; each fibre presents an appearance exactly like that of its neighbours, i.e., the 
muscle cells are all undergoing the same change equally and at the same rate. There is no appear- 
ance of one group degenerating, while another group is developing. 

(b) Osmic acid. No fat granules are seen anywhere. 

(c) Picric acid shows no karyokinesis in the muscle cells. The cells with the yellow granules 
are particularly numerous, especially in the submucous tissue. 

VII. 14 Days Post Partum. 

The muscle cells seem to be merely diminished in size, while their protoplasm (what remains 
of it) is less granular. 

VIII. 21, 28, and 36 Days Post Partum. 

Here uteri show much the same appearance as the 14 days' uterus. There is a very slight in- 
crease in the size of the muscle cell and its nucleus. 



So much for the description of the individual preparations. We take up now the 
conclusions as regards the muscular and connective tissues, and the blood-vessels. 

Conclusions as regards the Uterus in the Puerperium. 
1. Muscular Tissue. 

1. Degeneration. — The prime question that presents itself is — Does a degeneration of the 
muscle cells occur ; if so, what is its nature, in what position does it occur, and in what amount, 
i.e., does it attack all or some fibres, and at what time in the puerperium does it occur ? 



372 MR T. ARTHUR HELME ON HISTOLOGICAL OBSERVATIONS ON 

(a) Character of Change. It is evident that no such fatty degeneration as has been described 
by Kilian really occurs in the uterus undergoing the normal involution. I have been unable to find 
one single muscle cell whose substance gave the characteristic appearances or reactions of fat. 

On the other hand, the muscle cells do undergo a change ; but this change, though retrogressive, 
partakes more of the characters of an atrophy than a degeneration, the cell substance becoming 
.slightly dimmer and more granular, and progressively diminishing in amount. But it must be 
pointed out that this physiological retrogression differs totally from a pathological atrophy — in the 
latter it is the specific elements of a tissue that diminish and disappear, while the connective tissue 
Framework is not affected, or, if affected, it is in the way of increase and not of decrease, e.g., in the 
atrophy of muscle following destruction of its nerve supply there is found a diminution of the 
muscle cell substance, while the connective tissue shows an increase. On the other hand, in the 
physiological retrogression of the involving uterus, there is no increase of connective tissue, but an 
actual destruction and disappearance of it; while, so far as the muscle is concerned, merely a diminu- 
tion of the volume of its individual cells. The term granular atrophy would mislead, as suggesting 
an increase of connective tissue, and I prefer not to use it. The change is, therefore, one of physio- 
logical retrogression : the actual chemical character is not revealed by the microscope, but it is 
probable that the cell protoplasm undergoes a sort of peptonisation — all the surplus material that is 
no longer required being changed into some more soluble substance, which finds its way into the 
surrounding lymph — but, whatever its chemical nature, it is certainly not fatty, no traces of fat 
being found in the muscle cells, intercellular substance, or lymph spaces. 

(b) Time. As already described, when dealing with the muscle cells during pregnancy, towards 
full time the muscle fibres become greatly swollen and distended-looking, with a remarkably clear 
translucent or glassy appearance, evidently what is described as hyaline, while in many fibres spots 
which look like " vacuoles " have appeared, resembling the so-called serous degeneration. Now, 
within the first 24 hours of the puerperium, a tremendous change is wrought in these fibres, their 
size has been greatly diminished, and many of the fibres have already lost that clear glassy look, and 
instead present a dim or hazy, though still hyaline or translucent, refractile aspect (somewhat like 
ground-glass or spermaceti). Other fibres (especially in the vascular layer between the two muscular 
coats), greatly reduced in volume, still present the clear hyaline or glassy appearance of pregnancy, 
but by the second day all the fibres present the dim hyaline aspect ; and from this time onwards 
the diminution advances progressively and equally in all fibres alike till the 10th day, when it 
has reached its greatest height. There are evidently two stages in this post-partum change : — 
(1) An exceedingly rapid and remarkable alteration in appearance within the first 36 hours, i.e., 
a change from the clear hyaline to the dim hyaline appearance, accompanied by a rapid diminution 
in volume. (2) A gradual and progressive diminution in the volume of the dim hyaline fibres up to 
the 10th day. 

(c) Position. The diminution of volume occurs simultaneously in all fibres in all parts of the 
uterus ; it does not commence at one edge and advance, inwards or outwards, nor does it go on more 
rapidly at the inner or outer edge, but commences at the same time, and proceeds at the same rate 
in all the fibres throughout the organ. 

In the uterus 24 hours post partum one finds many fibres which have not yet lost their clear 
glassy appearance, and these especially around the vascular layer which separates the circular from 
the longitudinal muscular coat. 

All fibres, however, both clear and dim, are diminished in size. After 36 hours all the fibres 
present the same appearance, and henceforward undergo a parallel process of change, so that in a 
given uterus one finds each fibre in a condition similar to that of its neighbours. One does not find 
a group of degenerating cells lying by the side of a group of healthy cells, and another group of 
completely degenerated or young developing cells. 

(d) Amount of Diminution. Kilian,* in speaking of the rabbit's uterus, says- — " The involution, 
however, is not of the nature thought by obstetricians, but a complete atrophy, dissolving 

* " Die Structur des Uterus bei Thieren," von Dr F. M. Kiuan, Zeit.f. ration. Med., ix. Bd. 1 Hf't., 1849. 



THE MUSCULAR FIBRE AND CONNECTIVE TISSUE OF THE UTERUS. 373 

(Auflosung), and a new formation of young tissue in place of the old dead fibres," so that " a 
female that has gone through pregnancy, and the puerperal state, possesses at the end of the 
puerperium a new uterus." 

Heschl,* in reference to the human, says — " Of the uterus (which existed previous to labour) not 
a single fibre remains behind." 

My observations lead to a quite opposite conclusion, viz., this change does not bring about a 
destruction of the cells. Is is merely a transition from larger to smaller dimensions. There is no 
destruction, no diminution in number, but simply the removal of some of their contents, a diminu- 
tion in volume. It is an incomplete atrophy — not a numerical, but a simple physiological atrophy: 
the process in each fibre goes on to a certain point, and then stops — the fibre remains, only it is of 
a smaller volume. 

Under the view that the pregnant uterus at term was composed of two sets of muscle cells — 
those that existed before conception occurred (greatly augmented during pregnancy) and those that 
were newly developed during pregnancy — it became a question as to what fibres were attacked by 
degeneration during the puerperium. Two distinct views are held : — (1) Kilian and Heschl (the 
former basing his views on observations upon the rabbit's, the latter on the human, uterus) asserted 
that all fibres were destroyed. (2) Others (e.g., KoLLiKER) offered the theory, that possibly the original 
fibres were destroyed, and those of new development during pregnancy remained behind. 

This question is really settled by the observation that during pregnancy there is no new develop- 
ment of young fibres, consequently the uterus at term is made up of fibres which are, so to speak, 
contemporaries, and these fibres, which have all partaken in the great augmentation in volume of 
pregnancy, now all undergo an equal diminution of volume in the puerperium. This change occurs 
in all the muscle cells. 

2. Regeneration. — Prima facie, this complete destruction of one uterus, and the evolution of a 
new organ, seems in the highest degree, and on clinical grounds alone, improbable ; for, as is well 
known, the uterus that has once been pregnant remains after the puerperium of different form and 
larger than the virgin uterus. And the above observations on the changes in the muscle cells show 
that there is no need for new fibres to be formed. However, my preparations demonstrate the fact 
that no new fibres are developed from pre-existing muscle cells — i.e., no nuclear figures are to be 
discovered in any of the muscle cells. 

2. Connective Tissue. 

Concurrently with the changes in the muscle, there is also a change in the connective tissue, 
but the processes differ. Contrary to the experiences of Meola, I can find no such rapid transition 
of the connective tissue from a young to a riper state during the first days of the puerperium ; on 
the other hand, my preparations show a decided gradual and progressive diminution in amount and 
disappearance of the connective tissue. The process begins as an increasing granularity of the con- 
nective tissue fibres and cells immediately post partum. Some of the connective tissue fibres during 
the first three days swell up and become hyaline, eventually, however, breaking down into granules. 
The connective tissue cells also become more and more granular, their nuclei are greatly changed, 
breaking up into fragments which take on an intense colour when stained. In some cells there 
also appear very fine fat granules — this especially in the sub-peritoneal connective tissue, but also in 
some of the gland cells and sub-epithelial connective tissue cells also. In some places the destruc- 
tion is so complete that all traces of connective tissue fibres are lost, their place being taken by a 
clear mass, entangling granules and degenerating leucocytes and corpuscles and nuclei. Scattered 
through this also are the large plasmodia. Other large connective tissue cells do not undergo this 
degeneration. They are the large granular epithelioid cells with a single clear vesicular nucleus ; 
already by the first day they are seen lying round the blocked capillaries, but on the 6th day they 
are found to be almost filled with bright yellow granules derived from the disintegrating red blood- 

* " Untersuchungen iiber das Verhalten des menschlichen Uterus nach der Geburt," Zeit. der. Icais. kon. Gesellschaft 
der Aerzte zu Wien, Jahrgang 8, Bd. ii. pp. 228, &c, 1852. 



374 MR T. ARTHUR HELME ON HISTOLOGICAL OBSERVATIONS ON 

corpuscles. Their function is evidently to eat up and carry off the blood-pigment. They are 
especially numerous at the 10th day in the sub-epithelial connective tissue. After this time they 
diminish greatly in number, probably wandering off, but representatives are to be found months 
later. 

The Giant Cells. 

But of paramount importance and interest are those large multinuclear cells, which I have 
found in the full-time uterus, and described as plasmodia. These cells are present only during the 
last few days of pregnancy and the first days of the puerperium. What is their life-history ? 

Since the time when Haeckel first demonstrated the ingestion of colouring matter by the blood- 
corpuscles of various invertebrates, numerous researches have followed, showing that this is a 
property common to the wandering mesoderm cells, not only of invertebrates, but also of vertebrates, 
including man. Metschnikoff and many others have shown that these cells are capable of not 
only taking up foreign bodies introduced from without, but also resorbing structures that have no 
further use, e.g., in the atrophy of larval tissues during metamorphosis. 

Further, Metschnikoff has shown that these cells have the power of coming together and fusing 
to form one common mass or plasmodium, and that the so-called giant cells, so often found lying 
round foreign bodies, have " in all cases (in invertebrates) arisen by fusion of separate cells." 

Geddes and others have observed that even outside the body, and loitlwut the presence of foreign 
bodies, these cells have a great tendency to form plasmodia. 

Further, to quote Metschnikoff's own words — " In higher forms the mesoderm does not lose 
its primitive powers, but employs them against useless and harmful bodies, so that it retains the 
intracellular digestion, as well as many other of the characters of the protozoa — not only the power 
of throwing out pseudopodia, but also that of forming plasmodia. Mesodermal plasmodia are found 
even in the higher animals, not excepting man himself.""* 

Now it seems to me that here we have the explanation of what occurs in the rabbit's uterus. 

During the last days of pregnancy large clear cells with a single large nucleus appear in 
numbers (derived most probably from leucocytes) ; these coalesce and form the multinuclear cells — 
the plasmodia — so that by the end of pregnancy there is developed a large army ready for service. 
Immediately the uterus has emptied itself of its contents, and its greatly increased tissues 
completed their function, the plasmodia begin their work. They are no longer found lying in 
groups, but are scattered and their protoplasm becomes progressively more granular. Evidently 
their function is to eat up the waste material lying around them — whether in the form of granules 
from the connective tissue, or in solution from the muscle cells. 

The process might be compared to some extent with the resorption of cartilage by 
chondroclasts. 

In the developing bone we find a temporary framework of calcified cartilage, which is gradually 
eaten up by the chondroclasts, to make way for the permanent bone. 

In the uterus we find a somewhat analogous process. In the developing uterus of pregnancy 
we find a temporary framework of connective tissue developed to support the enlarged muscle cells, 
and to convey the greatly augmented capillary vessels for their nourishment. In the puerperal 
uterus this framework is no longer required, accordingly it is removed, and it is in this process that 
the plasmodia seem especially to be concerned. 

It may be well here to recall the fact, that other uninuclear wandering cells (phagocytes) 
assume the duty of ingesting the red blood-corpuscles, which lie in numbers in the blocked capillary 
vessels. 

As to the destiny of the plasmodia, after taking in their load they wander off from the uterus (none 
are to be seen on the 6th day post partum), and probably find their way into the general circulation 
— whether as plasmodia or after breaking up again into their original elements is not evident. 

* " Researches on the Intracellular Digestion of Invertebrates," by Dr Elias Metschnikoff, Quart. Jour. Micr. Sci., 
1884, vol. xxiv. p. 109. 



THE MUSCULAR FIBRE AND CONNECTIVE TISSUE OF THE UTERUS. 375 

3. Vessels and Blood-Corpuscles. 

In the earlier part of this paper I have alluded to the marked changes which occur in the vessels 
during the latest stages of pregnancy, and also to the importance of the sinus-like character of the veins. 
In the veins we find the following changes : — 

(1) An immediate compression of the sinus-like and smaller veins. 

(2) Many of these vessels later again become pervious ; but others remain permanently 
compressed, their endothelial cells coming to present a hyaline and granular appearance, gradually 
diminishing in volume, and finally disappearing. 

(3) In some of the veins a true process of proliferation of the intima goes on — this being started 
at points where the neighbouring endothelial cells are brought into contact by the folding of the 
vein wall on itself. 

In the arteries we find three varieties of change : — 

(1) Simple hyaline and granular change in the endothelial cells of the smaller arteries — many 
of which come to be compressed in the same way as the veins. 

(2) In the somewhat larger vessels, there is during pregnancy a hyaline swelling of the mus- 
cular and internal coats, both muscle cells and endothelium; this is followed by a gradual 
diminution in volume — just as in the muscle cells of the uterine wall. Their lumen is blocked 
from within by the swelling of the muscular coat and intima, instead of by compression from 
without. These vessels also gradually disappear. 

(3) In the larger arteries a true process of proliferative endarteritis (as shown in fig. 10) 
goes on. A rapid growth of cells takes place apparently from the endothelial cells and the sub- 
endothelial connective tissue of the intima ; the growth being generally asymmetrical — one side 
of the vessel wall growing more rapidly than the other, so that the lumen, instead of being circular, 
comes to be crescentic. In many cases the lumen becomes completely blocked. At the same 
time, the cells of the muscle coat become smaller, and probably eventually disappear. The lumen 
comes to be represented by a fibrous cord, which in its turn is gradually absorbed. 

It may be well to point out again the fact that the plasmodia are especially congregated 
round the vessels, and no doubt they assist in absorbing the contents of the cells of the vessels 
which are undergoing the gradual diminution. 

What is the fate of the red blood-corpuscles which lie in the blocked vessels and in the extra- 
vasations ? This is twofold : — 

(a) Many of the red corpuscles shrink and break up into a brownish granular mass — especially 
those which have escaped into the surrounding tissues — owing to rupture of the vessel wall. 

(b) The second process is that of absorption by cells — either the whole corpuscle being 
absorbed, or the above disintegrated fragments. Large clear endothelioid cells may be seen filled 
with brownish-yellow granules, and here and there whole corpuscles with a complete circular out- 
line. After absorption, the hsematin undergoes transformation into the brown pigment, the 
corpuscles breaking up into minute granules. 

I think there can be no doubt that this transformation of the haematin can go on both within 
and without the cells (phagocytes), because one sees — 

(1) "Whole corpuscles in the interior of cells. 

(2) Eed corps breaking down while still contained within the lumen of the vessel. 

Most of the phagocytes (which are especially numerous about the sixth day) wander off out of 
the uterus with their load of pigmented granules ; others, however, appear to remain permanently, 
for in the uterus which has once undergone the changes of the pregnancy and puerperium, there are 
always to be found a certain number of these cells with brown-yellow contents. 

Taking then a general retrospect of the changes that occur in the uterus from the virgin con- 
dition through pregnancy to the end of the puerperium, we may look upon the musculature as 
the essential or specific element of the uterus with a supporting framework of connective tissue. 

VOL. XXXV. PART 8. 3 Q 



376 MR T. ARTHUR HELME ON OBSERVATIONS OF THE UTERUS. 

The changes in the musculature may be compared to a wave. During pregnancy we have the 
ascent, the diminutive fibres becoming colossal in their dimensions ; then the descent of the puer- 
perium, the fibres diminishing again to their previous volume. To support this temporarily 
augmented musculature, there is a great development of a temporary connective tissue scaffolding. 
When the function of this augmented musculature is discharged, there is no further need of the 
support, which is accordingly destroyed and removed. 

Thus in the life-history of the uterine musculature we see, not as is generally thought, a fresh 
development of tissue to meet the new demand, to be followed later by a total destruction of the 
organ and its renovation or redevelopment, but a structure whose latent powers are evolved when 
occasion demands — a gradual rising to meet the needs of the economy followed by a corresponding 
return to the previous condition. 

The greater part of this work was carried out in the Pathological Institute of the University 
of Strassburg, the remainder in the Research Laboratory of the Royal College of Physicians, 
Edinburgh. 

I desire to express my best thanks to Professor von Recklinghausen and his assistant Dr 
Stilling, for their many kindnesses, and to the committee of the Physicians' Laboratory and Dr 
Woodhead, the superintendent, for their kind assistance. 



EXPLANATION OF PLATE. 

Dividing nucleus of endothelial cell of blood capillary, uterus of rabbit. 3rd day pregnancy, x 1200. 
Plasmodia from muscular coat of uterus of rabbit. 27th day pregnancy. x 350. 
Do. do. do. 26th day pregnancy. x 350. 

Do. do. do. 1st day puerperium. x 350. 

Plasmodia from mucous membrane of uterus of rabbit. 3rd day puerperium. x 350. 
6. Large periarterial venous sinus, from muscular coat of uterus of rabbit. x 350. 

Large granular plasmodial cell, with processes from muscular coat. 1st day puerperium. x 400. 
Plasmodia in muscular wall of uterus. 3rd day puerperium. x 350. 
9. Mucous membrane of uterus of rabbit, showing nuclei of gland epithelium in process of division ; also 
phagocytes containing yellow pigment. 3rd day puerperium. x 350. 
Proliferation of intima in an artery, from muscular coat of uterus. 10th day puerperium. 
Muscle cells of uterus, showing various stages in diminution of volume during the puerperium. 
1 and 2, 36 hours after labour. 

3, 3 days „ 

4 and 5, 6 days „ 

6 and 7, 10 days „ 

Fig. 1 2a. Isolated cells from subglandular connective tissue (mucous membrane) of uterus. These are prob- 
ably the cells which become "phagocytes." 36 hours after labour. 
Fig. 1 26 Isolated phagocytes containing yellow granules and red blood discs. 6 days after labour. 



Fig. 


1. 


Fig. 


2. 


Fig. 


3. 


Fig. 


4. 


Fig. 


5. 


Fig. 


6. 


Fig. 


7. 


Fig. 


8. 


Fig. 


9. 


Fig. 


10. 


Fig. 


11. 



Trans. Roy. Soc. Edin 1 ! -Vol. XXXV. 
T.A.Helme on the Muscular Fibre and Connective Tissue of the Uterus. 





£{$ SB W 









mm-. 




. ■ . 







e 









: fe, 



■ 



; mm 








■§Hm 



12.5 




M c F»]»: Jc Erstine L.th'f Edin' 



( 377 ) 



IX. — On Some Relations between Magnetism and Twist in Iron and Nickel. Part I. 
By Cargill G-. Knott, D.Sc. (Edin.), F.R.S.E., Professor of Physics, Imperial 
University, Tokyo, Japan. 

(Read 16th July 1888.) 

In a former paper # I described certain experiments on the relations of magnetism 
and twist in iron and nickel, the chief results of which it may be well to give briefly 
here. When an iron or nickel wire is under the influence of longitudinal and circular 
magnetisations, it twists in a direction which is definitely related to the direction of the 
magnetising forces. This effect in iron was discovered by Wiedemann^ and for conven- 
ience I shall call it the Wiedemann Effect. It was pointed out by Clerk Maxwell that 
the Wiedemann effect might be explained as a consequence of the earlier discovery made by 
Joule, that iron lengthens in the direction of magnetisation, and contracts at right angles 
thereto.^ Led by a consideration of Barrett's discovery § of the shortening of nickel wire 
in the direction of magnetisation, I determined to test nickel in the same way in which 
Wiedemann had tested iron. It was quite obvious that, if Maxwell's explanation of the 
Wiedemann effect were the true one, nickel wire should, ceteris paribus, twist in a sense 
opposite to that in which iron twists. The experiment when made completely fulfilled 
the expectation. Thus, when an iron wire, with one end fixed, is traversed by an electric 
current in the direction in which it is at the same time longitudinally magnetised, the 
wire is twisted so that the free end rotates right-handedly with reference to the traversing 
current, or the longitudinal magnetisation. In nickel, on the contrary, the corresponding- 
rotation is left-handed. This was the chief conclusion arrived at in my earlier paper ; 
and a little consideration will show how very readily the Wiedemann effect, whether in 
iron or in nickel, is explained in terms of the simpler strains studied by Joule and 
Barrett. 

There were, however, other results of interest touched upon, especially with regard to 
the influence of tension, which seemed to call for further investigation ; but it was not 
till the spring of 1887 that I was able to return again to the subject. In that year I 
was fortunate in obtaining the assistance of Mr Nagaoka, a graduating student of physics 
in the Imperial University of Japan, who undertook a very thorough examination of the 
influence of tension on the Wiedemann effect. His results, taken in conjunction with Mr 
Bid well's recent elaborate measurements of the changes of length of iron and nickel in 
varying magnetic fields, go far to establish the sufficiency of Maxwell's explanation, as 
will be seen further on. 

* " On Superposed Magnetisms in Iron and Nickel," Trans. Roy. Soc. Edin., vol. xxxii. p. 193, 1883. 

t Wiedemann's Galvanismus, Bd. ii. § 491 (1st edit.). 

t Sturgeon's Annals of Electricity, vol. viii. ; also Phil. Mag., 1847. 

§ See Nature, vol. xxvi., 1882. 

VOL. XXXV. PART 9. 3 R 



378 PROFESSOR KNOTT ON SOME RELATIONS BETWEEN 

The experiments were conducted in exactly the same manner as described in my 
earlier paper. The wire hung vertically within a magnetising solenoid, and to its lower 
and free end was attached a dipper, which dipped into a pool of mercury and made the 
necessary contact for the current along the wire. The tension in the wire was varied 
by putting on and taking off ordinary lead weights. The twist was measured by means 
of a mirror which was fixed to the free end of the wire, and reflected a spot of light in 
the customary manner upon a scale. 

When the wire was being traversed by a steady current so as to be circularly 
magnetised, the longitudinal magnetic field in which the wire was placed was repeatedly 
reversed until a steady difference of readings on the scale was obtained. The wire was, 
in other words, subjected to a cyclic straining, which gave a total to-and-fro twisting of 
the wire. For any one condition, some six or eight distinct successive readings were 
noted down, so that a good mean for the total cyclic twisting of the wire was obtained. 
Half the amount of this total twist was taken as the twist corresponding to the particular- 
combination of magnetising forces. 

With a given current along the wire, the current in the magnetising helix was taken 
through an ascending series of values. The highest field so. obtained usually lay between 
25 and 35 (C.G.S.) electromagnetic units. In the case of the iron wires, this greatest field 
was considerably higher than the field for which the maximum twist was obtained. In 
the case of the nickel wires, there was no such maximum twist observed up to the fields 
employed.* This point of maximum twist is one which offers distinct facilities for the 
discussion of the problem in hand, namely, the effect of tension on the phenomenon of 
twist. Does change of tension cause a change in the position of this maximum twist, and 
does it cause a change in its amount ? 

Some 120 distinct curves were obtained for different iron wires, showing the relation 
between twist and field for different currents along each wire and for different tensions. 
If we compare the curves for any one steady current along any one wire, we shall be able 
to study the direct influence of tension on the Wiedemann effect. The first conclusion is, 
that there is no evident relation between the tension and the position of the maximum; or, 
more .accurately, that the field, which for given current along the wire corresponds to the 
maximum twist, is in no way affected by change of tension through a considerable range. 
On the other hand, as was also clearly shown in my earlier paper, it is abundantly evident 
that the position of the maximum twist depends on the strength of the current along the 
wire — the stronger this current, the higher the field needed to produce the maximum 
twist. 

The effect, however, of tension upon the amount of twist is very marked. Thus, if 
we take any iron wire, and subject it at different tensions to the same combination of 
circular and longitudinal magnetic stresses, we shall find that the twist due to this com- 

* In more recent experiments, conducted on a somewhat different plan, I have heen able to obtain a maximum 
twist in nickel for intermediate values of field ; but much higher fields must be used than were available in Mr 
Nagaoka's experiments. 



MAGNETISM AND TWIST IN IRON AND NICKEL. 



379 



bination is always smaller for the higher tension. An increase of tension from 80 to 300 
kilogrammes weight per square centimetre is sufficient in some cases to diminish the twist 
due to a given combination of magnetisms to one-fourth of its original amount. In the 
following tables only the maximum twists are given, as it hardly seemed necessary to 
reproduce all the individual observations. We can safely employ these maximum twists, 
because of the first conclusion given above, namely, the absence of any recognisable 
relation between the tension and the longitudinal field corresponding to the maximum 
twist. 

In these tables the symbols have the following meanings : — 

r = radius of wire in centimetres. 

C = current along the wire in amperes. 

H = mean value of longitudinal magnetic field corresponding to the maximum 
twist. 

H' = C/5r = circular magnetic field at the circumference of the wire. 

T = tension in kilogrammes weight per square centimetre. 

6 = angular displacement, measured in radians, of any cross-section of the wire 
relatively to the cross-section 1 cm. distant from it. 









Table I 


— Iron 


Wires. 








No. 1. 


r=02. 


No. 2 


{continued,). 


. No. 2 


{continued). 




T 


104-0 




T 


104-0 




T 


10- 4 e 


C = 1-14 


196 


4-60 




543 


2-81 




1630 


1-97 


H=10-5 


594 


425 




698 


25 




1785 


1-75 


H'=11'4 


992 


335 




853 


225 




1940 


1-69 




1388 


255 




1087 


197 




2093 


1-56 




1784 


235 




1164 


1-88 




2256 


1-56 




2181 


1-60 




1312 


190 




2463 


1-22 




2577 


1-55 




1630 


1-47 




2618 


116 




2975 


115 




1785 
2093 


134 
119 




2773 

1087 

77 


103 

2-69 
3-84 


C = 173 


196 


500 




2256 


110 




H = 14-0 
H' = l7-3 


594 
992 


435 
400 




2463 
2618 


•90 

■84 














1388 


325 




2773 


•81 


No. 3. 


r=-04I5. 




1784 


300 




1087 


206 














2181 


235 




77 


350 




T 


104-0 




2577 


1-90 
















C = 152 
H=ll-8 


77 
232 


3-94 
3-78 


C =1-08 
H=7l 


48 
243 


2-96 
243 






No. 2. 


r = -032. 


H'= 95 


388 


3-69 


H' = 52 


439 


2-02 








543 


334 




633 


1-76 










T 


lO 4 -0 




698 


300 




864 


1-28 










853 

1087 


2-88 
269 




1058 


1-18 


C = -95 


77 


375 






1251 


104 


H =10-4 


232 


3-44 




1164 


2-44 




1413 


•92 


H'= 5-9 


388 


314 




1312 


219 




1661 


•67 



380 



PROFESSOR KNOTT ON SOME RELATIONS BETWEEN 



Table I. — Iron Wires — continued. 



No. 3 (continued). 


No. 4 (continued). 


No. 5 (continued). 




T 


io 4 -e 




T 


io 4 -e 




T 


lO 4 -0 




2050 
2241 
2505 


•58 
•53 
•53 




814 
1097 
1345 
1625 

1876 


170 
1-60 

1-22 

1-06 

•96 


C = 2-47 
H=113 
H'= 8-37 


23 

403 

784 

1164 

1551 

1931 


2-58 
213 
1-78 
1-39 
108 
•93 


C =1-6 

H =8-9 
H' = 77 


48 

243 

439 

633 

864 

1058 

1251 

1413 

1661 

2050 

2241 

2505 

2894 


316 

2-80 

2-39 

217 

1-69 

1-52 

1-33 

1-20 

101 

•77 

•65 

•68 

•53 


C = 2-07 
H=120 
H'= 8-25 


31 

284 
561 
814 
1097 
1345 
1625 
1876 


243 
2-33 
215 
1-91 
1-70 
1-46 
1-30 
112 


No. 6. r = -072. 




T 


io 4 -e 


C = 2-28 
H=10 
H'= 6-33 


20 

280 
536 

794 


1-79 
1-59 
1-39 
112 


No. 5. r = -059. 


C = 3-04 
H=116 
H'= 8-44 


15 

270 

527 
784 


1-88 
1-68 
1-46 
1-22 


No. 4. r=0502. 




T 


1O 4 '0 


C =152 
H=87 
H' = 515 


23 

41 

591 

797 

981 

1180 


232 
1-78 
1-39 
1-86 
117 
102 




T 


1O 4 "0 




C = 15 
H=123 
H'= 6 


31 

284 
561 


221 
213 
1-91 









A careful study of these numbers seems to lead to the following conclusions : — 
1. Other things being the same, the twist is greater in the thinner wire. It should 
be noted here that there is considerable difficulty in deciding as to the meaning of " other 
things being the same." The best mode is clearly, not to take the current through the 
wire as a guide, nor yet the current density, but something which may be regarded as 
giving an approximate estimate of the magnetic effect of the current. I have therefore 
taken the quantity H', which measures the direct magnetic force at the circumference of 
the wire due to an axial current of the magnitude used. The method is certainly open 
to criticism ; but in our absolute ignorance of the magnetic distribution in an iron wire 
due to a current passing along it, any other approximation could hardly be so justifiable. 
It will be shown later, that, on the simplest supposition possible, a calculation based on 
Maxwell's explanation of the Wiedemann effect leads to the result that the twist produced 
in a thin tube under the influence of given longitudinal and circular magnetising forces 
is inversely as the radius. It will readily be granted, when all the conditions are taken 
into account, that the experimental result just given is in fair accordance with the result 
deduced from theory. 



MAGNETISM AND TWIST IN IRON AND NICKEL. 381 

2. The longitudinal field necessary to produce the maximum twist is greater for the 
greater current along the wire. This result also may be shown to be in harmony with 
Maxwell's explanation. 

3. For a given combination of magnetising forces, the twist diminishes steadily as the 
tension increases. This relation also holds for all the other combinations of magnetising 
forces used in the experiments. The conclusion is in remarkable agreement with the 
results obtained by Mr Bid well in his elaborate investigations into the changes of length 
of iron in magnetic fields. Amongst other results, he found that under increasing loads 
the elongation of iron wire due to moderate magnetising forces decreases.* Here again 
Maxwell's explanation of the Wiedemann effect, in terms of the simpler Joule effect, 
accords well with the facts — both being powerfully influenced by tension, and that in the 
same direction. 

At first sight, it might be supposed that the maximum twist shown to exist in these 
experiments was exactly the same phenomenon as the maximum elongation of iron, 
obtained by Mr Bidwell in his experiments. But in trying to connect the two pecu- 
liarities, we encounter discrepancies which are hard to explain. Thus, the field which 
corresponds to the maximum elongation is much higher than the field which corresponds 
to the maximum twist. The latter, of course, depends on the strength of the current 
along the wire ; but in no case (out of nearly 200 distinct experiments with different 
wires) has the magnetic field corresponding to the maximum twist been higher than 25 
electromagnetic units — usually considerably lower, as in the table given above. Accord- 
ing to Mr Bid well's experiments, the field producing maximum elongation in a wire 1 '2 
mm. in diameter, varied from 45 for low loads to about 20 for high loads. This very 
striking effect of increased tension upon the strength of field required to produce the 
maximum elongation has, further, no analogue in the experiments on the maximum 
twist. Again, as bearing on this point, it may be mentioned that, although it is possible 
to obtain a maximum twist in nickel, there is no evidence of a maximum contraction. 
Hence the existence of the maximum twist does not imply the existence of maximum 
elongation or contraction, but clearly depends on other considerations. As I hope to 
show later, however, these considerations seem to be necessarily involved in the complete 
statement of the Maxwell explanation. 

We shall now pass to the discussion of the results for the nickel wires. Here we are 
not able to make use of such a well-marked singular point as a maximum, since no 
maximum was obtained with the magnetic fields employed. The main purpose of the 
present inquiry is, however, to consider the effect of tension on twist, other things being 
the same. It suffices, therefore, to fix upon some one value of the longitudinal field, which 
is common to all the experiments. In the following tables, only those twists are given 
which correspond to the field 28 "5. In the great majority of cases this was in reality 
one of the fields employed ; and in the comparatively few instances where it was not so, 
it was an easy matter to obtain by a simple interpolation an accurate enough value for 

* See Proc. Roy. Soc, vol. xl. p. 262 (1886). 



:382 



PROFESSOR KNOTT ON SOME RELATIONS BETWEEN 



the required twist. Only three different specimens of nickel wire were at our disposal, 
and of these the thinnest one had no special claims to purity. The symbols, r, c, H, H 7 , 
T, 0, have the same meanings as before, except that H has no reference whatever to 
a maximum twist. 

r = radius of wire in centimetres. 

C = current along the wire in amperes. 

H = longitudinal magnetic field. 

H' = C/5r = circular magnetic field at the circumference of the wire. 

T = tension in kgs. weight per sq. cm. 

6 == measure in radians of the twist per centimetre length of the wire. 

Table II. — Nickel Wires. 



No. 1. 



r = -0253. 



C = 152 
H =28-5 
H' = 12-0 



T 



160 

657 

1155 

1652 

2150 



10 4 -fl 



No. 2. 



r = -0422. 



521 

3-67 
301 

2-43 
1-82 



G = 2-28 


160 


6-32 


H =285 


657 


3-87 


H' = 18-0 


1145 


241 




1652 


2-22 




2150 


1-73 




2687 


1-85 



C = 152 
H=285 
H'= 72 



io 4 e 



No. 3. r=05. 



57-4 
236 
415 
594 

773 



1-54 
1-37 
1-25 
1-19 
107 



C = 2-28 


57-4 


1-51 


H=285 


236 


1-45 


H' = 108 


415 


1-45 




594 


1-38 



C = 152 
H = 285 
H'= 61 



41 

168 
296 
423 
550 
736 
1107 



C = 231 
H=28-5 
H'= 9-2 



41 
168 
296 
423 
550 
736 
1107 



1O 4 '0 



1-86 
2-31 
256 
2-64 
2-47 
235 
1-86 



301 
335 
3 38 
3-28 
315 
2-67 
210 



It will be noticed that the thinnest specimen behaves very like iron — that is, 
increasing tension is accompanied by diminishing twist, and that very markedly. A 
very similar effect is produced in the case of the intermediate specimen, with the single 
difference that the effect is not so pronounced. With the thickest, and what is probably 
the purest specimen, however, the effect of tension on the twist is quite peculiar. At 
first, as the tension is increased, the twist increases until the tension attains a value of 
300 or 400 kgs. weight per sq. cm. After this, as the tension is further increased, a 
pretty rapid decrease of twist sets in. It may be mentioned that this maximum twist 
for some intermediate value of tension was obtained in my earlier experiments * with 
very thin nickel wire. It existed, however, only in one series of experiments, and 
disappeared when the current along the wire was doubled. In the present case there is 



* See page 203 of the paper already referred to. 



MAGNETISM AND TWIST IN IRON AND NICKEL. 383 

a hint as to the manner in which the current along the wire influences the phenomenon. 
It would appear, in short, that the stronger the current along the wire the lower is the 
tension which produces the maximum twist in a given longitudinal field. Hence, for 
currents of considerable strength along the wire, it is quite possible that this maximum 
twist, occurring at a tension lower than the lowest used, could not be observed. For it 
must be remembered that in these experiments with a hanging wire, it is impossible to 
begin at zero tension, since the dipping arrangement needed for making contact and other 
necessary additions must have a definite weight. 

Thus we see that when nickel wire twists under the influence of circular and longi- 
tudinal magnetising forces, the amount of twist in certain specimens is influenced by 
tension in a manner very similar to what occurs in the case of iron — namely, the twist 
diminishes as the tension increases. In other specimens, however, a maximum twist is 
obtained for a certain intermediate tension ; and the tension which in a given longi- 
tudinal magnetic field corresponds to the maximum twist appears to be smaller for 
higher values of the circularly magnetising current along the wire. 

I am not aware that any experiments have been made upon the effect of tension on 
the contraction of nickel in longitudinal magnetic fields. Mr Bidwell does not seem to 
have studied the phenomena in nickel with anything like the thoroughness with which 
he has worked out the phenomena in iron. All I can find is the statement that " a 
nickel wire stretched by a weight undergoes retraction when magnetised," but whether 
this retraction is greater or less than the retraction in the unstretched case is not 
mentioned.* 

The tensions to which these various iron and nickel wires were subjected were 
obviously carried beyond the approximate limits of so-called perfect elasticity. A very 
natural inquiry to make was as to the effect of permanent strain upon the amount of 
twist for a given combination of magnetising forces. In the experiments on the iron 
wire of radius *032 (No. 2 in Table I.), observations were made as the tension was 
diminished again to its first and lowest value. The numbers will be found in the table, 
being in each case the last two rows of figures in the column. When the tension is 
reduced to its original value, there seems to be a slight decrease in the twist ; but this 
does not seem to hold for the intermediate conditions. All that can be safely said is 
that the effect of the permanent strain, after the stress is removed, is hardly appreciable 
in the case of iron. 

In the case of nickel, however, the effect of permanent strain is very marked, as the 
following small table will show. In this table two pairs of columns of tensions and 
twists are given. In the first of these are entered the tensions and twists before the 
wire had been subjected to a high tension of above 6000 kgs. weight per sq. cm.; 
and in the second column are entered the tensions and twists after the wire had been 
subjected to this high tension and relieved. 

* See Proc. Roy. Soc, vol. xl. p. 133. 



38-4 



PROFESSOR KNOTT ON SOME RELATIONS BETWEEN 



Before Strain. 


After Strain. 


T 


lO*-0 


T 


1O 4 


160 

657 

1155 

1652 

2150 


543 
3-68 
1-79 
1-88 ? 
103 


206 

847 

1487 

2121 

4150 


059 
•40 
•43 
•43 
•19 



These twists are produced by the combination of a longitudinal field of 2 7 "4 electro- 
magnetic (C.G.S.) units, with a current along the wire of 1*52 amperes. The radius of 
the wire, originally '0253 cm., was after the strain permanently reduced to '0223 cm. 

It is at once apparent that the stretched wire twists much less than the unstretched 
wire. The marked diminution in the radius is, of course, a sufficient indication of the 
great molecular change which stretching has produced in the wire. The wire has no 
doubt been considerably hardened by the process, and is no longer to be regarded as the 
same material. It should be mentioned that the numbers just given are only samples of 
the observations taken. A comparison of the twists in other fields than the one chosen 
leads, however, to the same general conclusion. 

It remains, finally, to consider the influence of change of temperature on the Wiede- 
mann effect. In order to carry out such an experiment, it was necessary to coil the 
magnetising helix upon a double-walled tube, between the concentric walls of which 
steam or other vapour could be passed. The upper end of the bore through which the 
wire hung was plugged up with cotton wool, so that no current of air could pass up or 
down. Under these conditions, the temperature of the wire may be assumed to be not 
very different from the temperature of the vapour after that has been for some time passed 
through the space within the walls. In the experiments as conducted, steam at 100° C. 
and water at 11° C. were passed in succession through the double- walled tube; and 
observations on the twist made in the usual way. The following table (Table III.) gives 
a few specimen numbers for both iron and nickel. 

As before, C is the current along the wire, and H 7 the corresponding magnetic force 
at the circumference of the wire. H is the longitudinal field. H and IT are measured 
in electromagnetic (C.G.S.) units. 

A glance at these numbers shows that the general tendency is for the twist to 
diminish as the temperature is raised, although at the highest tensions for iron there 
seems to be a tendency the other way. There is nothing as yet known regarding the 
influence of temperature on the simple Joule effect ; but we might safely argue from the 
results just given that in general change of length of iron and nickel in a given longi- 
tudinal field is greater at the lower temperature. 

In discussing these experiments, I have throughout assumed the truth of Maxwell's 
explanation of the Wiedemann effect in terms of the Joule effect. Wiedemann himself, 



MAGNETISM AND TWIST IN IRON AND NICKEL. 



385 



Table III. 



Iron. 


Nickel. 


r = "0415 cm. 
C — 16 amperes. 
H=10'5. 
H' = 77. 


r = "05 cm. 

C = 152 amperes. 

H=285. 

H'= 61. 


T 


10^6 


T 


io*-0 


11° 


100° 


11° 


100° 


48 

243 

439 

634 

828 

1030 

1260 

1450 


304 
2-65 
236 
207 
1-83 
1-76 
1-45 
106 


2-89 
2-60 
217 
1-93 
1-83 
1-64 
1-52 
1-33 


41 
168 
296 
423 
550 
736 


2-90 
294 

2-84 
2-80 
2-63 
268 


218 
239 
2-42 
250 
2-65 
255 



however, does not admit the sufficiency of this explanation. He entrenches himself 
behind the argument, that however neatly Maxwell's explanation may seem to explain 
the twist due to superposed magnetisms, it takes no cognizance of the reciprocal 
phenomena (see Beiblatter, 1886, vol. x. p. 728). Professor J. J. Thomson has shown 
(see his book Applications of Dynamics to Physics and Chemistry) that the existence 
of the twist produced by passing a current along a magnetised wire requires that when a 
current is passed along a twisted wire, or when a wire conveying a current is twisted, the 
wire becomes magnetised. This result is deduced simply from the application of 
recognised dynamical principles, and in no way takes account of any possible explanation 
of either phenomenon in terms of simpler ones. There is, so to speak, no stepping behind 
the scenes. It may well be doubted, however, if, after all, the experiments in which a 
twisted wire conveying a current is found to become magnetised, can be regarded as 
showing phenomena reciprocal to those in the experiments on the twisting of a magnetised 
wire under the influence of a current passing along it. For in the latter the twist pro- 
duced in the wire by the superposed magnetisms is always very small, far within the 
limits of torsional elasticity; whereas in the former it is necessary to apply comparatively 
large twists before any pronounced magnetic effect is obtained. In Professor Wiede- 
mann's own experiments the twists applied are very large indeed for the length of wire 
used, amounting to 7° per centimetre — that is, more than 200 times any of the twists 
obtained in the experiments just discussed, and 40 times the largest twist which I have 
ever obtained in like experiments. I have further found, by direct experiment, that it 
requires an applied twist of 1° per centimetre to produce any pronounced magnetic 
polarity in a wire conveying a current of half an ampere ; so that it seems to be quite 

VOL. XXXV. PART 9. 3 S 



386 PROFESSOR KNOTT ON SOME RELATIONS BETWEEN 

beyond the power of direct experiment to investigate the real reciprocal phenomena con- 
nected with what has been called in this paper the Wiedemann effect. 

Some of the weaknesses of Professor Wiedemann's own theory of frictionally restrained 
or rotating molecules as applied to these phenomena, have been very effectively exposed 
by Mr Bid well.* There are, besides, several objections of a general character which 
might be urged. For example, the theory is incapable of predicting phenomena. 
Maxwell's explanation, however, taken in conjunction with Bakrett's discovery of the 
shortening of nickel, at once suggested that the Wiedemann effect in nickel was opposite to 
that in iron ; and so I found it to be. Again, the shortening of iron in high fields at once 
suggested that in such fields iron should twist the other way ; in other words, behave like 
nickel in all fields ; and this Mr Bidwell found to be the case. Arguing from Mr Bidwell's 
recent work, we may expect to find that the Wiedemann effect in cobalt is in low fields 
as in nickel, but becomes reversed in high fields. In fact, if we call that twist positive 
which is shown by iron in low and moderate fields, we may state the relations of the 
Wiedemann effect in the three magnetic metals thus : in iron it is positive in low fields and 
negative in high ; in cobalt, it is probably negative in low fields and positive in high ; in 
nickel, it is always negative. 

Then, again, it is impossible to apply to Wiedemann's theory any numerical test, 

whereas, as I now proceed to show, Maxwell's explanation enables us to institute a 

numerical comparison between the Joule and the Wiedemann effects. We know by 

experiment that both phenomena exist ; and that, whether the latter is to be explained 

in terms of the former or not, they must at any rate coexist in the form of experiment 

more particularly under discussion. Let us then consider the distribution of stress and 

strain in a cylindrical tube which, as a whole, is subjected to the three strains — uniform 

elongation in the direction of its length, uniform expansion in directions at right angles 

thereto, and a simple twist about the axis of the cylinder. Let this axis be the z axis, 

and let the twist 6 be taken right-handedly with reference to it. The x and y axes 

will then lie in a section of the cylinder. If 73 is the elongation in the direction of the 

cylinder, and <r the elongation in any direction perpendicular thereto, we may express 

the displacements £, 77, £ of a point originally at x, y, z (z very small) by means of the 

following formulas: — 

£= — 6yz+o-x\ 

r)=+6xz+a-yY (1) 

From these we get, in the usual way, 

dx dz dy 

dy dx dz y 

dz dy dx 

* Phil. Mag., September 1886. 



MAGNETISM AND TWIST IN IKON AND NICKEL. 387 

and hence the equation to the strain ellipsoid is 

(1 - 2o-)X 2 + (1 - 2a) Y 2 + (1 - 2?«r)Z 2 - 26xYZ + 2ByXZ = 1 . 

By choosing the y or Y axis along the line from the centre of the section to the point 
x y, we get this equation in the somewhat simpler but no less general form — 

(l-2o-)X 2 +(l-2o-)Y 2 +(l-2sr)Z 2 +2%XZ = l .... (2) 

It is evident the Y axis is a principal axis of the strain ellipsoid ; and in finding the 
others we may confine our attention to the plane XZ. 

Let X /a be the direction cosines of a principal axis in this plane, and r the correspond- 
ing radius, then 

- 2<7\ 2 - 2sr M 2 + 2dy\ft = \-l 

\ 2 + M 2 = l. 

Differentiating and remembering that 1/r 2 is either a maximum or minimum, we get, on 

reduction, 

X_ By _ 2n-p ,„ 

fx 2<r-p~ By ' ' w ' 

where 

, 1 

and satisfies the equation 

p 2 -2p(T*+a) + 4>zro— (By) 2 = 0, (4). 

Now let r x r 2 be the maximum and minimum values of r, then we may write 

r 1 2 = l + 2ar 1 r 2 2 = l + 2 ( r 1 



1-^ = 2^ 1-^ = 2^, 



where m^ and <r l are the principal elongations, and 2^ 2^ are the roots of equation (4). 
Hence 

4z<r 1 a- 1 = 4sr<r — (By) 2 
from which we find easily, 

2ar = w x + o- 1= b sJi^-a-.Y-iBy) 2 



2a =^+0-^ J(*-<r x T-(e y y 

The equation of the strain ellipsoid, referred to its own principal axes, may then be put 
in the form 



(l-2w 1 )X 2 +{l-(7«r 1 + o- 1 )± V(*i-<Ji) 8 -(fy){ 2Y +(l-2o- 1 )Z 2 = l . 
To find the angle which the major axis of this ellipsoid makes with the axis of z, we 



:N8 professor KNOTT on some relations between 

have to solve equation (3), putting for p its proper value for the maximum radius ; in 
this case 1w v Then if is the required angle, we get 

tane=-^= 6 -y—=+ f ^ . . . (5). 

And now let us make the assumption, plausible enough and certainly the simplest that 
can be made in the circumstances, that this direction of maximum elongation coincides 
with the direction of the resultant magnetising force, as it may be assumed to exist in the 
experiments which give the Wiedemann effect. It will be remembered that we are 
applying the calculation to a cylindrical tube, although so far as the problem is an 
"elasticity" one there is no necessity for such a limitation. By so confining our attention 
to a thin- walled tube, we are able to regard the equation £, = zsz as a sufficiently near 
approximation to what might reasonably be expected to hold good if the experiment 
giving the Wiedemann effect were made with a tube instead of a wire, the tube being 
under the influence of an axial current and of a uniform longitudinal field. The magnetic 
distribution in an iron or nickel wire due to a current passing along it is something of 
which we know absolutely nothing, so that to make a calculation based upon an assumed 
expression for £ involving x and y in some simple manageable way, would probably ill 
repay the extra labour. 

Eeturning then to the equation (5), let us take a and ft as the circularly and longi- 
tudinally magnetising forces. Then, assuming that the maximum elongation ts 1 takes 
place in the direction of the resultant magnetising force, we may put tan 6 = a/ ft, and 
hence by a simple reduction, 

6y a/3 

2(sr 1 -cr 1 )~a2+/3 2, 

Here y is the radius of the tube. Writing it r, we obtain finally.* 

6 r oM--^ (6) ' 

that is, the twist 0, which measures the Wiedemann effect, is given in terms of the 
magnetising forces a, /3, the radius of the tube r, and the Joule elongation zf v together 
with cr 1; the accompanying elongation at right angles to vj v Of o^ we have no direct 
measurement. Joule, however, found that iron longitudinally magnetised did not change 
appreciably in volume. This would make o^ = — &J2 for the moderate magnetising 
forces with which Joule worked. 

In comparing this formula with results of experiments as obtained till now, we must 
remember that in the experiment we are dealing with a wire circularly magnetised 
throughout its interior in a complicated and altogether unknown manner ; whereas in the 
expression just given we are dealing with a thin-walled tube. Nevertheless, it is easy 
to see that the formula does to a certain extent apply even to the wire. Thus the twist 

* This expression differs from the one given in my earlier paper (p. 198). That, however, was incompletely worked 
out with a too early assumption of the law connecting the elongation with the magnetising force. 



MAGNETISM AND TWIST IN IRON AND NICKEL. 389 

is greater for smaller values of r. It is positive when (m — a) is positive, as in iron in 
moderate fields; it is negative when (vs—cr) is negative, as in nickel throughout, and as 
in iron in high fields. The existence of a maximum twist for some intermediate value 
of either a or /3 (ft or a remaining constant) will depend upon the particular way in 
which {zs — cr) depends on the quantities a and /3. Let us suppose, for instance, that a, 
the circularly magnetising force, is constant, and that /3 is allowed to vary through a 
large range. This supposition is a near enough approximation to the case of a wire con- 
veying a steady current, and then longitudinally magnetised. Now, it is clear that even 
if (ts — a) is constant, there is a maximum value for the twist given by the condition 
|3 = a. A maximum value of (m — a) for intermediate values of the field is not then a 
necessary condition for the existence of a maximum twist. Hence, it is not surprising 
that the field at which the maximum twist occurs should not be the same as the field at 
which the maximum elongation occurs. The maximum twist may exist without any 
maximum elongation ; as for example in nickel, in which I have recently obtained a maxi- 
mum twist about a field of 200 or 300. According to Mr Bidwell's recent experiments 
(see Nature, July 1888), nickel goes on distinctly contracting in magnetic fields up to 
750, after which up to 1300 or higher the length remains apparently constant. 

If we look closely at Mr Bidwell's curves of elongation for iron in ascending magnetic 
fields, we see that at first the curve is concave upwards, then becoming convex it reaches 
a maximum, after which it proceeds nearly straight in a long slope down to, and finally 
below, the zero line. For a considerable range of field near the point of inflexion we may 
regard the elongation as a linear function, of the most general form, of the magnetising 
force ; and such an assumption gives a maximum twist, or rather the possibility of it, 
except in the very special case of simple proportionality of the elongation to the magnetic 
force. 

Let us now test the expression for 6 by a direct numerical calculation, taking for this 
purpose the numbers given in Table I. for wire No. 1. Here r= '02, a= 11*4, /3=10'5, 
and we may take {?>*! — o^) to be approximately "000004. With these values, we get 

6 = -00002, 

about 2^ times smaller than the observed value for the wire. 

A similar calculation can very easily be made for nickel. Now the contraction for 
nickel in magnetic fields is considerably greater than the expansion for iron; and yet the 
twist numbers given in Table III. are sensibly of the same magnitude as those in Table I. 
The reason of this, however, is not far to seek. For it will be noticed that the factor 
a)S/(a 2 -f /3 2 ) is, because of the greater inequality of a and /3, much smaller than in the 
case of iron. 

At first sight, this may not seem to be a very promising result ; but when all the 
circumstances of the case are borne in mind, it will, I think, be admitted that the result 
is really as satisfactory as we could reasonably expect. The calculated twist for the tube 
is, at any rate, of the same order of quantity as the observed twist for the wire. The 

VOL. XXXV. PART 9. 3 T 



390 PROF. KNOTT ON MAGNETISM AND TWIST IN IRON AND NICKEL, 

calculation is perhaps interesting as being, I believe, the first of its kind, namely, a 
numerical comparison of what are at first sight different phenomena depending on the 
relations of magnetic and mechanical stress and strain. The result of the comparison, in 
my opinion, demonstrates the sufficiency of Maxwell's explanation of the Wiedemann 
effect in terms of the simpler Joule effect. Thus Maxwell's explanation has, to a first 
and simple approximation, stood the test of numerical calculation; whereas it is impossible 
even to imagine how to begin in applying such a test to Wiedemann's theory. 

Another serious objection to Wiedemann's theory is, that it gives a so-called explana- 
tion of a particular kind of magnetic strain, but furnishes no insight into the mechanism 
of simpler magnetic strains. And then, again, it seems to me — in making this statement 
I may simply be showing how little I understand the mechanism of the frictionally 
restrained rotating molecules — but it seems to me that, according to Wiedemann's theory, 
the direction of twist in an iron or nickel wire should depend on the order in which the 
circular and longitudinal magnetising forces are applied to the wire. Thus, let the wire, 
hanging vertically, be magnetised with north pole downwards, and then let a current 
be passed down it. Then the originally vertically polarised surface molecule facing the 
spectator will tend to rotate like the hands of a watch. But, if the wire is first magnetised 
circularly by a current flowing down it, and then subjected to the influence of the longi- 
tudinal field, the originally horizontally polarised surface molecule will tend to rotate 
contrary to the hands of a watch. Now, it is difficult to see how such contrary tendencies 
can possibly cause a similarly directed twist. 

I have recently found by experiment that the amount of twist, due to a given com- 
bination of magnetising forces, does depend upon the order in which the forces are 
applied ; but. except in a very particular case, the direction of twist never does. As the 
experiments, however, are not quite completed, I reserve their discussion for a second 
paper. 



( 391 ) 



X. — On the Fossil Plants in the Ravenhead Collection in the Free Library and Museum, 
Liverpool. By Robert Kidston, F.R.S.E., F.G.S. (Plates I. and II.) 

(Read 16th July 1888.) 

INTRODUCTION. 

A copy of Mr Marrat's paper " On the Fossil Ferns in the Ravenhead Collection " # 
having come into my hands, I was led to visit the Liverpool Free Public Museum in 
1886, and again in 1887, with the object of examining this interesting collection, and 
while doing so I received every assistance from Mr Thomas J. Moore, the Curator, and 
Mr F. P. Marrat, to whom I take this opportunity of expressing my indebtedness for 
the many kindnesses I received while studying the Ravenhead plants. I have further 
the pleasure of acknowledging the privilege accorded me by the Museum Committee, 
through the kind offices of the Rev. H. H. Higgins, which allowed me to have a number 
of specimens sent to Stirling, where I could more advantageously examine them than in 
the Museum, where the literature of the subject is limited. 

The history of the Ravenhead Collection has already been given by the Rev. H. H. 
Higgins t and by Mr Marrat in his paper " On the Fossil Ferns in the Ravenhead 
Collection." I may, however, briefly say here, that the specimens were collected by the 
Rev. H. H. Higgins, chiefly from the shales associated with some upright stems of fossil 
trees immediately below the lower of the two Ravenhead Coals. A few specimens may 
also have been collected from the shales between the two Ravenhead Coals, or in the 
shales that overlie them, but these form a small proportion of the collection. 

The section from which the fossils were collected was exposed while making a cutting 
on the Huyton and St Helens Railway, which passed through the Middle Coal Measures, 
at Ravenhead near St Helens, South Lancashire. 

Mr G. H. Morton, F.G-.S., author of the Geology of the Country around Liverpool, 
has most kindly favoured me with the following sketch of the Geology of the South- 
West Lancashire Coal Measures. Mr Morton's intimate knowledge of the geology 
of this district, and his reliable work on the subject, are already known to geologists, 
and I feel confident that the geological description with which he has favoured me 
will add much to the value of this paper. 

The Coal Measures of South-West Lancashire rest on the Millstone Grit, which is 
exposed in three restricted areas, viz., — Knowsley Park, Grimshaw Delf, and Parbold and 

* Abstract of Proc. of the Liverpool Geol. Soc, 13th session, 1871-72, p. 97, 1872 (Plates). 

t Proc. Liverpool Nat. Field Club for the year 1870-1871, and Abstr. of Proc. of the Liverpool Geol. Soc, 13th session, 
1871-1872, p. 94, 1872. 

VOL. XXXV. PART II. (NO. 10). 3 U 



592 



MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 



the Harrock Hills, as shown on the maps of the Geological Survey. The general succession 
of the Coal Measures shows the following subdivisions in the neighbourhood of Prescot, 
St Helens, and Wigan : — 

Upper Coal Measures. 

Middle or Productive Coal Measures. 

Lower Coal Measures. 

The Lower Coal Measures contain a greater proportion of sandstone strata than the 
other two subdivisions, and form the highest and most conspicuous ground; the Middle 
or Productive Coal Measures contain nearly all the coal seams that are of sufficient 
thickness to be worth working ; while the Upper Coal Measures are of little economic 
importance. These subdivisions are principally the result of observations by Professor 
Edward Hull, F.R.S., who originally surveyed the district for the Geological Survey 
thirty years ago, though it had been described long before by the late Mr E. W. Binney, 
F.R.S., to whom geologists are so much indebted for his valuable and long-continued 
researches in the Lancashire Coal Field. The succession of the coal seams and the 
thickness of the strata between them have been known for a long period, this information 
having been gradually obtained from the sinking of shafts and working the coals during 
the last one hundred years, but some of the most reliable details are given in the 
" Sections of Shafts sunk in the Middle Coal Measures of Prescot, St Helens, Wigan, and 
Burnley," by Messrs C. E. de Rance, F.G.S., and A, Strahan, B.A., F.G.S., published 
by the Geological Survey. 



Railway Station. 



Prescot. 




Upper. 



Middle. Lower. 

Fig. 1.— Section through the Coal Measures at Prescot. 



The above section (fig. 1) along four miles of country from north to south, with 
Prescot in the centre, exhibits in regular sequence the nearest development of the Coal 
Measures in the country around Liverpool, and shows the subdivisions described. This 
threefold succession is common to the whole of the Lancashire Coal Field, which was all 
formed under the same general conditions, but it has been since broken up by faults, and 
denuded to such an extent that the subdivisions appear at the surface in separate areas, 
and the whole is frequently covered by the overlying Triassic strata which extend over a 
large portion of Lancashire and Cheshire. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 393 

Lower Coal Measures. 

These overlie the Millstone Grit, and form the lowest subdivision of the Coal 
Measures. They have been called the Gannister Series, a provincial term, alluding 
to the hard siliceous beds which often form the floor of the few coal seams they 
contain. The formerly well-known Huyton Quarry* was worked in strata in this 
subdivision, and the hard grey flagstones occur in other parts of the Coal Field, 
and form some of the highest beds, as at CJpholland, where they occur at the east end 
of the tunnel. 

All traces of the quarry have now been covered up, but forty years ago a good 
section of the strata was exposed, and 50 feet of grey micaceous flagstones, with some 
6 feet of shale near the top, formed an anticlinal arch parallel with the railway, with the 
beds dipping to the east and west. The surfaces of many of the beds were beautifully 
ripple-marked, and tracks of Annelids were abundant. Some curious casts, supposed 
to have been those of the burrows of bivalves, and Calamites Cistii, Brongt., were of 
common occurrence. The strata exposed in the quarry were more fully shown in the 
railway cutting from Huyton to St Helens, about 200 yards farther north, where the 
grey flagstones were found to be 60 feet thick, with 10 feet of shale and a thin bed of 
coal (1 foot 6 inches) over them. The beds were found to dip at various angles to the 
south-east, and included some that were worked below the bottom of the old quarry. 
Shales and flagstones, with the seams of coal known as " Mountain Mines," occur at a 
considerable depth below the Huyton flagstones, which are 240 feet below Little Delf 
Coal at the base of the Middle Coal Measures. 

The Lower Coal Measures are continued at the surface from Huyton Quarry to the 
east of Prescot over a large area about Hurst House and the Hazels. Another area, 
forming the eastern portion of Knowsley Park, belongs to the same subdivision. Near 
Hag Delf, in the Park, a little east of the Stand Quarry, a bed of coal (2 feet 4 inches), one 
of the "Mountain Mines," occurs, and is said to have been worked about a hundred years 
ago. The old pit banks remain, and traces of iron smelting occur at the same place ; but 
trees have grown over the spot, and little is known as to when the coal and ironstone 
were worked. The highest beds occur along the eastern boundary of the Park, par- 
ticularly about Trap Wood, where flagstones and shales have been exposed, though little 
can be seen of them now, and much less of the strata in the Park generally than when 
explored in 1862. 

There is a considerable area, several square miles, of the Lower Coal Measures 
between Prescot and Parbold ; and the Liverpool and Yorkshire Kail way presents a fine 
section of them at Pimbo Lane Station, which was originally described by Mr Binney, who 
found here Goniatites Listeri and Aviculopecten papyraceus, fossils characteristic of the 
horizon. The beds are exposed in a deep cutting, and consist of black shales, micaceous 

* A Station on the Liverpool and Manchester Railway. 



394 MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 

flaggy sandstones, and two or three thin beds of coal. To the south of Pimbo Lane is 
Billinge Hill, where there are quarries in the sandstone near the Beacon, which is on the 
highest ground, 593 feet above Ordnance datum. Probably, in consequence of the 
predominating sandstones, the subdivision forms an elevated district, and good sections of 
the strata are of frequent occurrence. In the Geology of Wig an* Professor Hull 
gives a section through the Lower Coal Measures at Billinge, which shows the thickness 
of the successive beds, and that they are altogether 1881 feet thick. There are six beds 
of coal, the Upper Mountain Mine being 2 feet, and the Loiver Mountain Mine being 2 
feet 8 inches thick, with four intermediate seams, much too thin to be worked. In a 
sandstone quarry near the top of Billinge Hill, Catamites J Bothrodendron, and Lepido- 
dendron have been found. 



Middle or Productive Coal Measures. 

The Middle Coal Measures form the most important subdivision, economically 
considered, for they contain all the valuable beds of coal. They extend from near 
Huyton and south of St Helens, many miles to the north-east, forming two projections, 
bounded on each side by the Trias, a few miles to the east of Liverpool. Although there 
are many quarries and small exposures where they may be seen, there is no important 
section showing a great thickness of the beds and the relative position of the various 
shales, sandstones, and coal seams. The whole of the strata between the Lyon's Del/ 
Coal at the top, and the Little Delf Coal or Arley Mine at the bottom, are considered 
to constitute the subdivision. 

The Middle Coal Measures at Prescot are the nearest to Liverpool, and may be adopted 
as a scale for comparison with those at St Helens and Wigan, and the coal seams at each 
of these places have been correlated with those on the same horizon, though they are 
usually known by different names, which were originally given to them before it was 
possible to ascertain their relative position in the Coal Field. The following General 
Section, compiled from the Geological Survey Memoirs, is the final result of observa- 
tions extending over many years. 

The section given on the opposite page shows that the Middle Coal Measures are 
thinner at Prescot than at St Helens, where the coal seams are at greater distances 
from each other, while at Wigan there is a still greater expansion of the series. It also 
shows the equivalent of each coal seam at Prescot, St Helens, and Wigan ; as, for 
example, that the Tenlands Coal at Prescot becomes the Ravenhead Coals at St Helens, 
and the Wigan 5-feet Mine at Wigan ; and that the Little Delf Coal at the two former 
places becomes the Arley Mine, or Orrell 4-feet at the latter place. The succession of 
the coal seams could only have been ascertained from the records obtained in sinking 

* Memoirs of the Geological Survey. 

t The Liverpool Museum contains specimens of Calamites Suckowii and Bothrodendron (decorticated) from this locality. 
-R. K. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 395 



General Section of the Coal Measures. 



Prescot. 


Thickness. 


St Helens. 


Thickness. 


Wigan. 


Thickness. 




Ft. 


In. 




Ft. 


Iu. 




Ft. In. 








Lyon's Delf coal, 


2 


8 


Riding Mine coal, . 


3 8 


Strata, . 








Strata, 


47 


4 


Strata, 


36 


Felcroft coal,* . 


7 





London Delf coal, . 


2 


6 


Ince Yard Mine, 


2 6 


Strata, . 


96 





Strata, 


86 


2 


Strata, 


108 


Pastures coal, . 


4 


6 


Potatoe Delf coal,* . 


5 


3 


Ince Four- Feet coal, 


3 6 


Strata, . 


30 





Strata, 


41 


9 


Strata, 


150 9 


Discoverer coal* 


3 





Earthy Delf coal * . 


6 


2 


Ince Seven-Feet coal, 


6 


Strata, . 


57 





Strata, 


65 





) 




Yard Mine coal, 


3 


2 


Coal, 


2 





\ Strata, 


71 5 


Strata, . 


42 





Strata, 


53 





J 




Cannel Mine coal * . 


5 


9 


Coal* . 


6 


4 


Wilcock, or Furness 
coal* 


4 7 


Strata, . 


75 





Strata, 


147 


8 


Strata, 


252 


Higher Bug coal, 


6 


6 


St Helens Main coal, 

Strata, 
Cannel coal, . 


9 
9 

2 




3 


I Pemberton Five- 
j* Feet Mine, 


5 2 


Strata, . 


6 





Strata, 


21 


3 


Strata, 


30 9 


Lower Bug coal, 


3 


6 


Four-Feet coal, 


3 


2 


Little coal, 


2 6 


Strata, . 


69 





Strata, 


56 





Strata, 


45 1 


Little End, or Cheshire 






Pigeon House coal, . 


2 





Pemberton Four- 




coal, 


2 


3 








Feet coal, 


4 6 


Strata, . 


252 





Strata, 


275 


7 


Strata, 


387 


r 






Ravenhead Higher 






"N 




Tenlands coal* . .< 


12 


6 


coal, 
Strata, 


3 

4 


10 
2 


1 Wigan Five - Feet 
Mine, 


5 








Ravenhead Main 






<. 






Delf coal, 


7 





_> 




Strata, . 


80 





Strata, 


73 





Strata, 


90 


Bastion's coal * . 


4 





Bastion's coal,* 


4 


3 


Wigan Four - Feet 
Mine, 


4 








( Strata, 


22 





Strata, 


72 


Strata, . 


100 





3 Higher Roger coal,* 
\ Strata, 


6 
123 


6 



Wigan Nine - Feet 
Mine* . 
) 


9 


Sir John coal * . 


3 


4 


Sir John coal, . 


3 


4 


> Strata, 


280 


Strata, . 


50 





Strata. 


57 





J 




Prescot Main coal,* . 


10 





Flaggy Delf coal, 


4 


8 


Cannel and King 
coals* 


6 








Strata, 


78 





Strata, 


66 








Lower Roger coal ,* . 


5 


3 


Ravin Mine* . 
Strata, 


3 
168 


Strata, . 


240 





Strata, 


333 


8 


Haigh Yard coal, . 

Strata, 
Bone coal, 

Strata, 


3 
150 

2 3 

7 8 


Rushy Park coal, 


5 





Rushy Park coal, 


4 


6 


Orrell Five-Feet coal, 


3 6 


Strata, . 


150 





Strata, 


162 





Strata, 


186 


Little Delf coal, 


3 





Little Delf coal, 


3 





Arley Mine, Orrell 
Four-Feet, 


4 


1320 


6 


1739 


3 


2172 10 



* Contain beds of shale, or coal of inferior quality. Coal seams under 2 feet in thickness are omitted, not being 
worth working. 



:*9G 



MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 



the shafts of pits, and from borings, though there are many places where the outcrops 
of beds of coal may be seen at the surface and the associated strata examined. 

In 1870 the Ravenhead Higher Coal and the Ravenhead Main Del/ Coal at Thatto 
Heath, were both exposed during the construction of the Huyton and St Helens 
Railway, though when it was finished the sides of the cutting were levelled, and traces 
of the coal seams almost obliterated. The following is a section of the strata (fig. 2), 
and the locality will always be of local interest on account of the numerous fossil plants 
that have been obtained there through the patient industry of the Rev. H. H. Higgins, 
who fortunately resided so near the place that he was able to visit it almost daily for 
several months. 




1. Pigeon House Coal. 2. Ravenhead Higher Coal. 3. Ravenhead Main Coal. 4. Bunter Pebble-Beds. 

5. Thatto Heath Sandstone. 6. Shales and Sandstone. 7. Fossil Trees. 

Fig. 2. — Coal Measures at Ravenhead, St Helens. (Scale, 12 inches to 1 mile in length.) 



The position of the coal seams is shown on the sketch (fig. 2), and that of a remark- 
able line of trunks of trees in the position in which they originally grew, usually about 
4 or 5 feet high from the roots upwards, and about 8 feet below the Eavenhead Main 
Coal. It was principally in the black shales and in ironstone nodules below the coal 
seam and below the trees that the plant remains were found, and the number of species 
collected within such a restricted space was extraordinary. The collection when com- 
pleted was presented by the Rev. H. H. Higgins to the Liverpool Free Public Library 
and Museum, where it is known as the " Ravenhead Collection." Two wings of an 
orthopterous insect, Protophasmidce, from the same beds as the plants, are also preserved 
in the collection. 

In the Geology of Wigan, Professor Hull describes several sections where coal 
seams might be seen cropping out at the surface, but that of the Ince 7 -feet Coal, in 
a cliff above Leyland Mill, nearly two miles north of the railway station, is the only 
one that is now well exposed, and may be seen from the road with a series of associated 
shales and sandstones. The Middle Coal Measures may be seen to advantage in a 
large quarry on the south-west of St Helens, where Messrs Doulton & Co. obtain 
clay for making bricks and tiles; the strata are here nearly 118 feet in thickness. 
The highest beds consist of sandstone, with a great series of shales and thin sand- 
stones below, and three thin seams of coal about the middle. They are all above the 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 397 

Rushy Park Coal, which is about 200 feet below the lowest beds visible. Plant remains 
are numerous in the shales, and on one side of the quarry two upright trunks of trees 
are exposed. 

In the Middle Coal Measures there are a few species of Mollusca, — Anthrocosia 
robusta being the most common ; they usually occur in thick bands and extend a con- 
siderable distance from one coal pit to another, particularly a band above and another 
below the Rushy Park Coal. Fish remains also occur in many localities, but direct 
search has not been made for them, though there is little doubt that the scales and teeth 
of small species would be found of common occurrence, and those of Ccelacanthus, 
Gyrolepis, and Platysomus have been already recorded from several localities. 



Upper Coal Measures. 

These consist of red and purple shales, clays, and sandstones when seen within 200 
or 300 feet from the surface, probably coloured by the infiltration of ferruginous water 
from the overlying Permian or Trias, but of black shales and grey sandstones where 
brought up during borings and the sinking of shafts. The first exposure is in the 
cutting of the Liverpool and Manchester Eailway, between Rainhill and Marshall's Cross 
stations, and the strata are principally of a red colour. They may occasionally be seen 
in small openings about the reservoirs south of St Helens, but as there are no coal 




Dip 10 



2" Strong brown stone. 

Soft crumbly metal. 
W 



Fig. 3. — Permian or Trias on Coal Measures, Haydock. (Scale, 1 inch to 6 feet.) 

seams of sufficient thickness to be worth working, and the strata of little economic 
importance, they are seldom exposed. At Collin's Green and Bold, shafts have been 
sunk through the base of the Trias and the whole of the Upper Coal Measures into the 
Middle Coal Measures beneath, and both have afforded accurate sections and proved the 
thickness of the strata at these two places, which are about a mile apart. Mr A. 
Strahan, in the Geology of Prescot (2nd edition), gives the details of each section, 
and the debris from the shafts still remains on the surface. The strata consist of 
various shales and sandstones, with a red stain for some distance below the surface, but 
as the depth increases they gradually present the usual black and grey coloured shales, 
In both the Collin's Green and Bold sections there are numerous thin beds of coal, 
varying from 1 to 24 inches in thickness. A shaft was sunk at Haydock a few 



398 MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 

years ago, passing through the Trias for 101 feet into the Upper Coal Measures, but did 
not reach any coal seams. At the end of a heading 327 feet from the surface, the 
Permian or Trias was found resting unconformably on the Upper Coal Measures, as 
shown in fig. 3. 

These important sections prove the thickness of the subdivision to be about 
1200 feet, and that in any future sinking or boring through the Trias for the purpose 
of obtaining coal, the occurrence of that amount of unproductive strata may be expected 
before the productive Coal Measures are reached. 



SYNOPSIS OF SPECIES* 

Calamarise. 

Calamites, Suckow. 

Group I. — Calami tin a, Weiss [Emend.), 1884, Steinkohlen Calamarien, 

part ii. p. 59.t 

Calamitina (Calamites) varians, Sternb., var. inconstans, Weiss, 

PL I. figs. 1, la. 

Calamites (Calamitina) varians inconstans, Weiss, Steinkohlen-Calamarien, part ii. pp. 62 and 69, pi. xvi.a, 

figs. 7, 8; pi. xxv. fig. 2, 1884. 
Calamites varians, Sternb., Vers., ii. p. 50, pi. xii. (cast). 
Cyclocladia major, Feistmantel (not L. & H. (1), (in part), Vers. d. bokm. Ablager., Abth. i. pi. i. 

fig. 8. 
Calamitina Gbpperti, Weiss, (not Ettingsnausen), Steinkohlen Calamarien, part i. pi. xvii. figs. 1, 2, 

1876. 

Description. — Internodes broader than long, faintly striate, periods of six to fourteen 
non-branch-bearing nodes between each vcrticel of branch-scars; branch-scars oval, 
approximate, with a more or less central circular cicatrice ; the internode at the base of 
the period is the shortest, the next succeeding internodes have generally a slight increase 
in their length, but sometimes their increase in length is irregular, the uppermost one is 
generally the longest ; leaf-scars contiguous or slightly distant, transversely oval or sub- 
triangular. 

Remarks. — The only specimen of this species in the collection is shown natural size 
on PI. I. fig. 1, and measures about 32 cm. in length. It shows two complete and 

* In the Ravenhead collection are several ferns which are specifically distinct from any included in this list, but 
are too fragmentary for any satisfactory determination. 

t In Abhandl. z. geol. special-harte v. PrevAsen u. Thiiringischen Staaten, Band v. part ii. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 399 

two incomplete periods. The lengths of the internodes entering into the composition 
of the specimen are as follow : — 



Period incomplete. 



Period complete. 



Period complete. < 



Period incomplete. ■{ 





mm. 


Internode 1 


8-50 


2 


5 


3 


6-50 


4 


5-50 


5 


4-50 


6 


5 


„ 7 


3 


8 


4 


9 


3-50 


„ 10 


3 


„ 11 


3 


„ 12 


7 


„ 13 


6 


„ 14 


8-50 


„ 15 


8 


„ 16 


8 


„ 17 


7-50 


„ 18 


7-50 


„ 19 


7 


„ 20 


5 


„ 21 


4-50 


„ 22 


4 


„ 23 


3-50 


„ 24 


7-50 


„ 25 


8 


„ 26 


11 


„ 27 


12 


„ 28 


12-50 


„ 29 


12 


„ 30 


11-50 


„ 31 


11 


„ 32 


10-50 


„ 33 


8-50 


„ 34 


9-50 


„ 35 


6-50 


„ 36 


5 


„ 37 


9-50 


„ 38 


19 


„ 39 


22-50 


„ 40 


17 



Node with branch-scars. 



Node with branch-scars. 



Node with branch-scars. 



321-50 mm. 



I am not certain that the full width of the specimen is shown at any part. On the 
right-hand side the true margin of the fossil is seen, on the left, from the manner in 
which the matrix has splintered off, it is difficult to decide as to whether the full 



VOL. XXXV. PART II. (NO. 10). 



3 x 



400 MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 

width of the stem is preserved, though it possibly may be at the two parts marked 
with a + . 

1 at first regarded this specimen as the type of a new species, chiefly on account of 
the subtriangular leaf-scars, but before raising it to that rank I forwarded a sketch of 
the plant to Dr Weiss, Berlin, who kindly examined it, and now informs me that he 
believes the " Eavenhead " specimen cannot be separated from Calamitina varians, var. 
mconstans. He further states that he has observed the subtriangular form of the leaf- 
scars on other specimens, and he believes it to be the result of pressure, which has 
brought about a slight distortion. 



Calamitina (Calamites) varians, Sternb., var. 

Catamites varians, Sternb., Vers., ii. p. 50. 

Calamitina (Calamites) approximatus, Brongt. 

Catamites approximatus, Brongt. (in part), Hist. d. veget. foss., p. 133, pi. xxiv. figs. 2, 3, 4, 5. 



Group II. — Eu calamites, Weiss, 1884, Steinkohlen Calamarien, 

part ii. p. 96. 

Eucalamites (Calamites) ramosus, Artis. 

Catamites ramosus, Artis, Antedil. Phyt., pi. ii. 

Calamites (Eucalamites) ramosus, Weiss, Steinkotden Calamarien, part ii. p. 98, pi. ii. fig. 3 ; v. figs. 1, 2 

vi.; vii. figs. 1, 2; viii. figs. 1, 2, 4; ix. figs. 1, 2; x. fig. 1; xx. figs. 1, 2. 
Calamites (Eucalamites) ramosus, Kidston, Trans. Roy. Soc. Edin., vol. xxxiii. p. 341. 
Calamites nodosus, L. & H., Fossil Flora, vol. i. pis. xv., xvi. 

As foliage: — 

Annularia radiata, Brongt., Prodrome, p. 156. 



Group III. — S tylocalamites, Weiss, 1884, Steinkohlen Calamarien, 

part ii. p. 119. 

Stylocalamites (Calamites) Suckowii, Brongt. 

Catamites Suckowii, Brongt., Hist. d. vcget. foss., p. 124 (pi. xiv. fig. 6?), pi. xv. figs. 1-6; pi. xvi. 

figs. 2, 3, 4 (fig. 1 ?). 
Calamites Suckowii, Geinitz, Vers. d. Steinkf. in Sadism., p. 6, pi. xiii. figs. 1-6. 
Calamites Suckowii, Weiss., Steinkotden Calamarien, part i. p. 123, pi. xix. fig. 1, 1876 ; part ii. p. 129, 

pi. ii. fig. 1; pi. iii. figs. 2, 3; pi. iv. fig. 1; pi. xxvii. fig. 3 (1884). 
Calamites Suckowii, Zeiller, Flore foss. du bassiri houiller de Valenciennes, p. 333, pi. liv. figs. 2, 3 ; 

pi. Iv. fig. 1. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 401 



Stylocalamites (Calamites) undulatus, Sternb. 

Catamites undulatus, Sternb., Vers., i. fasc. 4, p. xxvi. ; Vers., ii. p. 47, pi. i. fig. 1 (pi. xx. fig. 81). 

Calamites undulatus, Brongt., Hist. d. veget. foss., p. 127, pi. xvii. figs. 1-4. 

Calamites undulatus, Sauveur, Veget. foss. d. terr. houiller de la Belgique, pi. v. figs. 1-3 ; pi. viii. fig. 1. 

Calamites undulatus, Zeiller, Flore foss. du bassin houiller de Valenciennes, p. 338, pi. liv. figs. 1 and 4. 

(?) Calamites decoratus, Brongt. (in part), Hist. d. veget. foss., p. 123, pi. xiv. figs. 3, 4. 

Calamites (Stylocalamites) Suchowii, var. undulatus, Weiss, Steinkohlen Calamarien, part ii. pp. 129, 

134, 135, pi. xvii. fig. 4. 
Calamites cannaformis, Roehl (not Schlotheim) (in part), Foss. Flora d. Steink. Form. Westph., p. 12, 

pi. ii. fig. 3. (In Palaiont., vol. xviii.) 
Calamites inaiquus, Achepohl., Niederrh. Westfdl. Steinkohl., p. 114, pi. xxxiv. fig. 15. 
Calamites duplex, Achepohl., Niederrh. Westfdl Steinkohl., p. 135, pi. xli. fig. 11. 

Remarks. — I previously regarded Calamites undulatus, Sternb., as a varietal form — 
probably of Calamites SucJcowii, Brongt. — but now believe it to be a distinct species. 
Calamites undulatus is most easily distinguished from Calamites SucJcoivii by the ribs 
being undulate (but this character may be removed by pressure), and by their termin- 
ating at the nodes in sharp triangular points — not being rounded, as in Calamites 
Suckoivii. 

Stylocalamites (Calamites) Cistii, Brongt. 

Calamites Cistii, Brongt., Hist, de veget. foss., p. 129, pi. xx. 

Calamites Cistii, Zeiller, Flore foss. du bassin houiller de Valenciennes, p. 342, pi. lvi. figs. 1, 2. 

Calamocladus, Schimper. 
Calamocladus equisetiformis, Schloth., sp. 

Calamocladus equisetiformis, Schimper, Traite d. paleont. veget., vol. i. p. 324, pi. xxii. figs. 1-3. 
Asterophyllites equisetiformis, Zeiller, Flore foss. du bassin houiller de Valenciennes, p. 368, pi. lviii. 

figs. 1-7. 
Casuarinites equisetiformis, Schloth., Flora d. Vorwelt, p. 30, pi. i. figs. 1, 2 ; pi. ii. fig. 3. 

Calamocladus grandis, Sternb., sp. 

Calamocladus grandis, Schimper, Traite d. paleont. veget., vol. i. p. 325. 

Asterophyllites grandis, Zeiller, Flore foss. du bassin houiller de Valenciennes, p. 376, pi. lix. figs. 4-7. 

Bechera grandis, Sternb., Vers., i. fasc. 4, p. xxx. pi. xlix. fig. 1. 

Calamocladus lycopodioides, Zeiller, sp. 

Asterophyllites lycopodioides, Zeiller, Flore foss. du bassin houiller de Valenciennes, Atlas, pi. lix. figs. 1, 2 
(1886) ; Text, p. 380 (1888). 

Sphenophylle^e. 

Sphenophyllum, Brongt. 
Sphenophyllum cuneifolium, Sternb., sp. 

Sphenophyllum cuneifolium, Zeiller, Flore foss. du bassin houiller de Valenciennes, p. 413, pi. lxii. fig. 1 ; 

pi. lxiii. figs. 1-10 (includes varieties). 
Rotularia cuneifolia, Sternb., Vers., i. p. 33, pi. xxvi. figs. 4a, 46. 



402 MR EOBERT KIDSTON ON THE FOSSIL PLANTS IN THE 

FlLICACE^E. 

Sphyropteris, Stur, 1883. 

Sphyrqpteris, Stur, Morph. u. Syst. d. Culm. u. Carbon Fame, p. 23 (in lxxxviii. vol. Sitzb. d. k. Akad. 

d. Wissensch., 1 Abth.). 
Sphyropteris, Stur, Carbon Flora, vol. i. p. 16. 
Sphyropteris, Zeiller, Flore foss. d.u. bassin houiller de Valenciennes, p. 31, fig. 17. 

Description. — Sporangia free, exannulate, globular or hemispherical, placed in rows 
on oblique, levated bands, which are borne on a narrow membranous expansion, ter- 
minating the pinnae and pinnules, and placed at right angles to them. Sporangia walls 
composed of elongated cells, which diverge from a terminal depression (pore V). 

Remarks. — This peculiar genus is represented in the Eavenhead Collection by a 
single, but fortunately well-preserved specimen, of Sphyropteris obliqua, Marrat, sp. 
( = Sphyropteris Crepini, Stur). This example gives some additional information in 
regard to the structure of the sporangia. These are globular or perhaps hemispherical, 
as they seem to be attached by a base whose width is equal to the diameter of the 
sporangia, the cells of whose w r alls radiate from a central depression (PL I. figs. 3c and 
3d). This central depression is too constant to be accidental, and I can only explain 
it by supposing it to be an apical pore for the dissemination of the spores, similar to 
what occurs in the fossil genus Urnatopteris and in the recent genus Dancea. 

Sphyropteris obliqua, Marrat, sp. 

PL I. figs. 3, 3a, 3b, 3c, 3d. 

Sphenopteris obliqua, Marrat, Proc. Liverpool Geol. Soc, Session 13, 1871-72, p. 99, pi. ix. fig. 3, 1872. 
Sphyropteris Crepini, Stur, Morph. u. Syst. d. Culm. u. Carbon Fame, p. 24, fig. 6c, 1883. 
Sphyropteris Crepini, Stur, Carbon Flora, vol. i. p. 18, pi. xxxix. figs. 1, la.; text, fig. 6c, p. 16, 1885. 

Description. — Frond tripinnate, pinna alternate, lanceolate ; pinnules broadly oval, 
uppermost almost simple or slightly dentate, lower divided into 4-6 truncate lobes, which 
are occasionally notched ; veins indistinct, but apparently one enters each lobe. 
Sporangia as described above, and placed on obliquely elevated bands, which are situated 
on linear membranous expansions placed at right angles to the extremities of the 
pinnules or terminating the pinnae. 

Remarks. — PI. I. fig. 3a, gives an enlarged view of an ultimate pinna, magnified five 
times. Two of the pinnules and the apex of the pinna each bear one of the linear sporangial 
expansions. At fig. 3b, one of these sporangial bands is shown enlarged 7\ times. The 
left half of the band was not so well preserved as the right half, and on it the sporangia 
were a good deal displaced, but on the right and better preserved portion three elevated 
ridges are clearly i=jhown, on two of which are borne three sporangia each, and on the 
terminal one two have been developed. Their number may, however, vary. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 403 

There is no recent genus with which Sphyropteris can be compared in the arrange- 
ment of its sporangia and the peculiarly developed sporangia! band. In the structure 
of the individual sporangia, however, they are clearly shown to be Marattiaceous, and 
in their apparently opening by a terminal pore, they show some similarity to the 
sporangia of the recent Danwa. 

Sphyropteris Crepini, Stur, does not appear to differ in any character from 
Sphyropteris obliqua, Marrat, sp. 

Zeilleria, Kidston. 
Zeilleria delicatula, Sternb., sp. 

Zeilleria delicatula, Kidston, Quart. Jour. Geol. Soc, vol. xl. p. 592, pi. xxv. 

Zeilleria delicatula, Kidston, Catal. Palaioz. Plants, p. 66, 

Sphenopteris delicatula, Sternb., Vers., i. fasc. ii. p. 30, pi. xxvi. fig. 5 ; fasc. iv. p. xvi. 

Sphenopteris, Brongt. 
Sphenopteris Sauveurii, Crepin. 

Sphenopteris Sauveurii, Crepin, Bull. Soc. Roy. Bot. de Belgique, vol. xix. part ii. p. 17. 

Sphenopteris Sauveurii, Crepin, ibid., vol. xx. part ii. p. 26. 

Sphenopteris Sauveurii, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 79, pi. ix. fig. 6.* 

Sphenopteris trifoliolata, Artis, sp. 

Sphenopteris trifoliolata, Brongt., Prodrome, p. 51. 
Filicites trifoliolatus, Artis, Antedil. Phyt., pi. xi. 

Remarh. — For notes on Sphenopteris trifoliolata, see the following species. 

Sphenopteris Marratii, Kidston, n. sp., 
PI. II. figs. 1, 2. 

Sphenopteris-Gymnogramrnides, Proc Liverpool Nat. Field Club for 1870-71, frontispiece, fig. 1, 1871. 
Sphenopderis trifoliolata, Marrat (not Artis), Proc. Liverpool Geol. Soc, Session 13, 1871-72, p. 98, 1872. 

Description. — Frond bi- or tripinnate, pinnse linear lanceolate, alternate ; pinnules 
small, of firm texture, opposite or alternate, obovate, obcordate, or trilobate ; the pinnules 
towards the base of the pinnse stalked, the uppermost pinnules sessile or attached by a 
wide pedicle-like contraction of the pinnule. Main rachis of primary (?) pinnae or 
frond (?) moderately thick, the other rachis thin. Venation not shown. 

Remarks. — There is no more difficult class of fossils to determine accurately than 
the small pinnuled Sphenopterids, and it is only after a good deal of consideration that I 
have come to the conclusion that a new specific name must be applied to the two 
specimens given on PL II. figs. 1, 2. 

Sphenopteris Marratii appears to me to be altogether different from Sphenopteris 

* A fall synonymy is given here by M. Zeiller of this much-confused and many -named species. 



404 MR ROBERT K1DSTON ON THE FOSSIL PLANTS IN THE 

trifoliolata, Artis, sp., with which it had been previously identified. In this latter 
species the pinnules are much larger and characteristically trilobate, and the plant has a 
different mode of growth, — the importance of these characters, though difficult to describe 
adequately in words, are very obvious when actual specimens of the two plants are 
compared. The type figure of Sph. trifoliolata * is a very characteristic drawing of the 
plant, and shows the differences more clearly than can be pointed out in a written 
description. 

In PI. II. fig. 1, the pinnae have an almost linear outline, while in Sphenopteris 
trifoliolata they are more deltoid. Some of the pinnules of Sphenopteris Marratii, 
especially on the upper pinnae, are trilobate (fig. lb), but on the lower pinnae, where the 
trilobate pinnules are characteristic in Sph. trifoliolata, they do not occur on Sph. 
Marratii, or only very rarely. 

Fig. 2 is from another small specimen of the same species, and shows portions of 
two pinnae, evidently belonging to a lower part of the frond than that from which fig. 1 
has been derived. It is laxer in its growth, the ultimate pinnae and pinnules are more 
distant, but still possess the same characters as those of fig. 1. 

The pinnules at the base of the pinnae are often notched or obcordate, those higher on 
the pinnae obovate. The terminal pinnules are blunt, and frequently lobed. The number 
of pinnules on the ultimate pinnae varies according to the position of the pinnae on the 
frond. 

Fig. 1 is probably the terminal portion of a frond, though it may be only the 
termination of a primary lateral pinna. 

Sphenopteris Marratii may possibly be the same plant as that figured by Stur as 
Sphenopteris trifoliolata,^ but in the absence of enlarged drawings of the pinnules of his 
fern, one cannot be certain of their identity. One point, however, is certain, that the 
fern named Diplothmema trifoliolatum, Artis, sp., by Stur, is not that plant. 

Of the references Stur gives for Sphenopteris trifoliolata, Artis, sp., only that to 
Artis' figure is correct. The Sphenopteris numularia figured by Andr^,| and the 
Sphenopteris trifoliolata given by Brongniart,§ not being referable to Artis' species. 
In fact, Sphen. trifoliolata, Artis, sp., seems to be very much misunderstood on the 
Continent. 

Since preparing the catalogue of Palaeozoic Plants in the British Museum, I have had 
considerable opportunities of examining specimens of Sphen, trifoliolata, and now think 
that none of the figures issued under this name by Andr^e,|| Roehl,1T or Brongniart ** 
belong to that plant. 

* Antedil. Phyt., pi. xi. 

t Carbon Flora, p. 346, pi. xix. figs. 1-4 (Diplothmema trifoliolatum). 
Vorwelt Pflanzen, pi. xi. 

§ Hist. d. ve'gd. foss., pi. liii. fig. 3. 

|| Vorwelt Pflanzen, pi. ix. figs. 2-4. 

IT Foss. Flora d. Steink. Form. Westph., pi. xvi. figs. 3 and 16. 
** Hist. d. ve'ge't. foss., pi. liii. fig. 3. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 405 

Zeiller, in his admirable work on the Flore foss. d. bassin houill. d. Valenciennes, 
gives, on his pi. i. figs. 1-4, drawings of a fern he identifies as Sphenopteris trifoliolata, 
Artis, sp. None of these figures are, however, according to my view, referable to the 
true Sph. trifoliolata, but to Sph. obtusiloba. Though the pinnules of some of Zeiller' s 
figures show a trilobate segmentation, they are much more solid, and altogether want 
the light and diffuse appearance of the pinnules of Sph. trifoliolata. I also fear that 
the union of Sph. numularia with Sph. trifoliolata, as proposed by Zeiller, is inad- 
missible. I am led to this conclusion from the examination of specimens of Sph. 
trifoliolata from the same horizon as that from which the type was procured, and 
therefore do not think that I can be mistaken in the characters of the true plant. 

Sphenopteris Marratii approaches very closely to Sphenopteris (Diplothmema) 
avoldensis, Stur, # but seems to differ from it in the pinnae and pinnules being more 
obtuse, and according to the description of Sphen. avoldensis, Stur, sp., it must have 
been of more delicate texture. Dr Stur here again gives no enlarged drawings of the 
segmentation of his species, without which a satisfactory comparison of these small 
pinnuled Sphenopterids is almost impossible. 

I have pleasure in naming the Eavenhead species after Mr Marrat, who was the first 
to publish a list of the Ravenhead ferns. 

Sphenopteris obtusiloba, Brongt. 

Sphenopteris obtusiloba, Brongt., Hist. d. veget. foss., p. 204, pi. liii. fig. 2. 

Splieiwpteris obtusiloba, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 65, pi. iii. figs. 1-4; pi. iv. 

fig. 1 ; pi. v. figs. 1, 2 (exclude syn. Sph. grandifrons, Sauv.). 
Sphenopteris irregularis, Sternb., Vers., ii. p. 63, pi. xvii. fig. 4. 
Sphenopteris trifoliolata, Zeiller (not Artis), Flore foss. d. bassin houill. d. Valenciennes, p. 75, pi. i. 

figs. 1-4. (Ref. in part.) 

Remarks. — See notes appended to Sphenopteris trifoliolata. 

Sphenopteris mixta, Schimper. 

Sphenopteris mixta, Schimper, Traite d. paleont. veget., vol. i. p. 382, 1869. 

Sphenopteris mixta, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 95, pi. xii. fig. 3, 1886. 
Sphenopteris mixta, Lesquereux, Geol. Survey of Illin., vol. iv. p. 409, pi. xv. fig. 7, 1870. 
Sphenopteris mixta, Lesquereux, Coal Flora, vol. i. p. 276 (pi. liv. figs. 1-3 (?). 

Sphenopteris rigida (?), Lesquereux (not Brongt.), Geol. Survey of Illin., vol. ii. p. 435, pi. xxxix. figs. 5, 6. 
Sphenopteris (Aneimioides) (?) pulchra, Marrat, Proc. Liverpool Geol. Soc, Session 13, 1871-72, p. 101, 
pi. viii. fig. 1. 

Remarks. — In the Ravenhead Collection is a small specimen which I refer to this species. 
It appears to be similar to the fern figured under this name by Zeiller in his Flore 
foss. d. bassin houill. d. Valenciennes. In the description of his specimen he mentions 
that the inferior surface of the pinnules is entirely covered with adpressed hairs. A 
similar structure occurs on the Ravenhead example, but is on the upper surface of the 
pinnules. 

* Carbon Flora, vol. i. p. 345, pi. xxiv. tig. 6. 



406 MR ROBERT KTDSTON ON THE FOSSIL PLANTS IN THE 

Mr Lacoe has kindly sent me a specimen of Sph. mixta, labelled as from the 
" Sub-carboniferous of Clinton, Mo." This series is below the Millstone Grit, and there- 
fore from a much lower horizon than that from which the British example originates. 
1 have compared our plant most carefully with Mr Lacoe's specimen of Sphenopteris 
mixta, and in nervation and pinnule cutting it agrees entirely, and although the example 
sent me by Mr Lacoe does not show the adpressed hairs, yet the surface of the pinnules 
have a roughened appearance, suggestive of the presence of hairs. Lesquereux, in the 
original description of Sph. mixta, which he then identified as Sphen. rigida, Brongt. (?), 
says, "in our specimens the surface of the leaves is evidently rugose, as if it had been 
originally squamose or hairy."* In his latest description in the Coal Floral he says, 
" surface smooth or polished/' but at the same time gives his Sphen. rigida (Lesqx., not 
Brongt.) as a synonym to Sphen. mixta. Zeiller thinks it probable that the fern 
Lesquereux figures in the Coal Flora may not belong to Sphen. mixta, Schimper, but 
be a distinct species. On this question, in the absence of American specimens similar to 
those figured in the Coal Flora, I cannot enter. 

The Lancashire fossil has been named Sphenopteris pulchra by Mr Marrat in his 
list of the Ravenhead Ferns, but it does not differ in any point from Sphenopteris mixta, 
Schimper, as far as I understand this species. 

Sphenopteris coriacea, Marrat. 

PI. I. figs. 4, 4a, 4b. 

Sphenopteris coriacea, Marrat, Proc. Liverpool Geol. Soc, Session 13, 1871-72, p. 98, pi. ix. figs. 1-2, 

1872. 

Description. — Frond ; pinnae alternate, ovate-lanceolate, main rachis 

thick ; pinnules crowded on lower pinnse, more distant on upper ones, contracted at the 
base, deltoid or lanceolate, coriaceous, basal pinnules divided into oblong lobes, the basal 
lobe much developed and directed towards the base of the pinna, and again divided into 
oblong simple or lobed segments ; in these pinnules the basal lobe is not much less than 
the other part of the pinnule. Upper pinnules oblong, and divided into 5-9 simple or 
lobed teeth; terminal lobe blunt, and not enlarged. Veins immersed, obscure. 

Sphenopteris Footneri, Marrat. 
PL II. figs. 3, 3a, 36. 

Sphenopteris Footneri, Marrat, Proc. Liverpool Geol. Soc, Session 13, 1871-72, p. 101, pi. viii. figs. 2,3, 

1872. 

Description. — Frond tripinnate or decompound ; rachis winged, flexuous ; pinnae 
alternate, narrow deltoid ; pinnules oblong, of delicate texture, and divided into 3-7 
blunt segments, these in turn bear 2-4 blunt teeth, into each of which extends a vein ; 

* Geol. Survey of Illin., vol. iv. p. 436. t Page 276. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 407 

the terminal lobe of the pinnule usually consists of two blunt equal lobes separated by a 
sinus. The pinnules become confluent on the upper pinnae, and on the upper portion of 
the lower pinnse. Veins distinct, flexuous. 

Remarks. — This description is founded on the specimen figured here, — one of Mr 
Maerat's types, — and on another specimen in my collection from the Middle Coal 
Measures, Claycross, Derbyshire, given me by Dr Pegler, Stonebroom. 

Fig. 3 shows portion of a pinna, probably from the basal part of a frond. My 
Claycross specimen seems to show a portion of a frond near the apex. In that figured 
the pinnules are free, and attached to a broad winged rachis ; in the Claycross example 
the pinnules are confluent on the greater number of the pinnse, like those on the apical 
portion of the pinna given at 3b. The venation is firm and distinct, and when the limb 
of the pinnule has been subjected to any decay before fossilisation, little more than the 
veins are preserved. 

This species is closely related to Sphenopteris gracilis, Brongt.,* from Newcastle-on- 
Tyne, especially as figured by ZEiLLER,t and will likely require to be united with it, but 
as probably at an early date I may have further facilities for examining additional 
specimens, I meantime retain Mr Marrat's name for the Ravenhead fern, though 
strongly of opinion that the plant should be referred to Sphenopteris {Renaultia) 
gracilis, Brongt. 

I have hitherto been unable to ascertain to which species Brongntart's Sphenopteris 
gracilis really referred, and it is only since seeing M. Zeiller's figures that it has 
occurred to me that the Sphenopteris Footneri, Marrat. is most probably the Sphen. 
gracilis of Brongniart. 

Sphenopteris spinosa, Goppert. 

Sphenopteris spinosa, Gbpp., Gatt. d. foss. Pflanzen, Lief 3-4, p. 70, pi. xii. 

Sphenopteris spinosa, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 135, pi. xv. figs. 1-3. 

Remarks. — The specimens are all very fragmentary. 

Sphenopteris furcata, Brongt. 

Sphenopteris furcata, Brongt., Hist. d. veget. foss., p. 179, pi. xlix. figs. 4, 5. 
Diplothmema furcatum, Stur, Carbon Flora, vol. i. p. 299, pi. xxviii. figs. 2, 3. 

Diplothmema furcatum, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 147, pi. iv. figs. 5, 6; pi. 
v. fig. 4. 

Spenopteris multifida, Lindley & Hutton. 

Sphenopteris midtifida, L. & H., Fossil Flora, vol. ii. pi. cxxiii. 

Remarks. — In the Quart. Jour. Geol. Soc, vol. xl. p. 594, and in my Catalogue 
of Palaeozoic Plants, p. 65, I have united this species with Urnatopteris tenella, Brongt., 

* Hist. d. ve'ge't. foss., p. 197, pi. liv. fig. 2. 

t Flore foss. d. bassin houill. d. Valenciennes, p. 94, pi. iv. figs. 2, 3. 

VOL. XXXV. PART II. (NO. 10). 3 Y 



408 MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 

sp. Though not yet quite certain as to the true characters of the plant which was 
named Splienopteris multijida by Lindley and Hutton. I believe that the fragments 
here recorded under this name are referable to that species ; and if I am correct in this 
supposition, Splienopteris multijida, L. & H., cannot be united with Urnatopteris tenella, 
Brongt., sp. 

Sphenopteris Sternbergii, Ett., sp. 

Asplenites Sternbergii, Ettingsbausen, Steinkf. v. Radnitz., p. 42, pi. xx. figs. 2, 3, and 4 (pars). 
Pecopteris Sternbergii, Schimper, Traite d. paleont. veget., vol. i. p. 525. 
Alethopteris Sternbergii, Kidston, Catal. Palceoz. Plants, p. 138. 

Splienopteris Sternbergii, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 128, pi. ix. fig. 5; pi. 
xxxviii. fig. 6. 

Neuropteris, Brongt. 
Neuropertis heterophylla, Brongt. 

Neuropteris heterophylla, Brongt., Hist. d. veget. foss., p. 243, pi. lxxi.; pi. lxxii. fig. 2. 

Neuropteris heterophylla, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 261, pis. xliii. and xliv. 

Neuropteris Loshii, Brongt., Hist. d. veget. foss., p. 242, pi. lxxii. fig. 1; pi. lxxiii. 

Cyclopteris trichomanoides, Brongt., Hist. d. veget. foss., p. 217, pi. lxi. bis, fig. 4. 

Neuropteris tenuifolia, Schlotheim, sp. 

Neuropteris tenuifolia, Brongt., Hist. d. veget. foss., p. 241, pi. lxxii. fig. 3. 

Neuropteris tenuifolia, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 273, pi. xlvi. fig. 1. 

Filicites tenuifolius, Schloth., Petrefactenkunde, p. 405, pi. xxii. fig. 1. 

Neuropteris gigantea, Sternberg. 

Neuropteris gigantea, Sternb., Vers., i. fasc. 4, p. xvi. 
Neuropteris gigantea, Brongt., Hist. d. veget. foss., p. 240, pi. lxix. 
Osmunda gigantea, Sternb., Vers., i. fasc. 2, pp. 33 and 36, pi. xxii. 

Neuropteris macrophylla, Brongt. 

Neuropteris macrophylla, Brongt., Hist. d. veget. foss., p. 235, pi. lxv. fig. 1. 

Neuropteris macrophylla, Kidston, Trans. Roy. Soc. Edin., vol. xxxiii. p. 354, pi. xxi. fig. 2 ; pi. xxii. 
figs. 2, 3. 

Neuropteris dentata, Lesquereux. 
PI. II. fig. 5. 

Neuropteris dentata, Lesqx., Gcol. of Pennsyl., 1858, p. 859, pi. v. figs. 9, 10. 
Neuropteris dentata, Lesqx., Coal Flora, voL i. p. 82, pi. v. figs. 7, 8. 
Nephropteris denliculata, Marrat, Proc. Liverpool Oeol. Soc, p. 104, 1871-72. 
(Unnamed), Higgins, Proc. Liverpool Nat. Field Club, frontispiece, fig. 13, for year 1871-72. 

Description. — Pinnules ovate, obtuse, truncate, or subcordate at the base, irregularly 
lacerate-dentate, the clefts occasionally extending some distance inwards from the 
margin. Veins dichotomous, flabellate, slightly arched in passing to the border, thin, 
close. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 409 

Remarks. — This species is only represented by a single pinnule, but it agrees so 
completely with Lesqueeeux's description and figure that I have no doubt that the 
Eavenhead specimen is identical with Lesqueeeux's species. 

Neuropteris dentata is one of a group of Neuropterids whose pinnules have dentate 
or ciliate margins, and whose relationship to each other is not yet clearly made out. They 
usually occur in a very fragmentary condition. 

In my paper on the Fossil Flora of the Radstock Series, I have given three figures of 
what I believe to be the Neuropteris Jimbriata, Lesqx.* 

Heee describes two species which are very closely related to these, viz., Neuropteris 
(Cychpteris) lacerata and Neuropteris (Cyclopteris) ciliata,f which perhaps should be 
referred to one or other of the American species. 

Odontopteris, Brongt. 
Odontopteris Reichiana, Gutbier. 

Odontopteris Reichiana, Gutbier, Vers. d. Zwick. Schwarzk., p. 65, pi. ix. figs. 1, 2, 3, 5, 7 ; pi. x. fig. 13. 
Odontopteris Reichiana, Weiss, Foss. Flora d. jiingst. Stk. u. Rothl., p. 32, pi. i. figs. 3-9. 
Odontopteris neuropteroides, Marrat (not Roemer, not Lesqx.), Proc. Liverpool Geol. Soc, Session 13, 
1871-72, p. 107, pi. vii. figs. 1, 2. 

? Odontopteris Britannica, Gutbier. 

Odontopteris Britannica, Gutbier, Vers. d. Zwick. Schwarzk., p. 68, pi. ix. figs. 8-11. 
Odontopteris Britannica, Weiss, Foss. Flora d. jiingst. Stk. u. Rothl., p. 45, pi. i. fig. 2. 

Mariopteris, Zeiller. 
Mariopteris muricata, Schloth., sp. 

Mariopteris muricata, Zeiller, Flore foss. d. bassin. houill d. Valenciennes, p. 173, pi. xx. figs. 1-4 ; 

pi. xxi. fig. 1; pi. xxii. figs. 1, 2; pi. xxiii. fig. 1 (includes var. nervosa). 
Pecopteris muricata, Brongt., Hist. d. veget. foss., p. 352, pi. xcv. figs. 3, 4; pi. xcvii. 
Filicites muricatus, Schlotb., Flora d. Vorwelt, pp. 52 and 55, pi. xii. figs. 21 and 23. 
Pecopteris nervosa, Brongt., Hist. d. veget. foss., p. 297, pi. xciv.j pi. xcv. figs. 1, 2. 

Remarks. — Specimens of the typical form and of Pec. nervosa, Brongt., which is only 
a variety of Mariopteris {Pecopteris) muricata, are contained in the collection. 

Pecopteris, Brongt. 
? Pecopteris Miltoni, Artis, sp. 

Pecopteris Miltoni, Kidston, Trans. Roy. Soc. Edin., vol. xxxiii. p. 374. 
Filicites Miltoni, Artis, Antedil. Phyt., pi. xiv. 

Dactylotheca, Zeiller. 
Dactylotheca plumosa, Artis, sp. 

Filiates plumosa, Artis, Antedil. Phyt., pi. xvii. 

* Trans. Roy. Soc. Edin., vol. xxxiii. p. 361, pi. xxi. figs. 3, 4, 1888. 
t Flora fossilis Helvetia:, p. 17, pi. vi. figs. 17 and 24. 



410 MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 

Remarks. — I have previously placed Dactylotheca dentata, Brongt., sp., with this fern 
as a variety of it, and this view may be correct ; but it must be noted that Dactylotheca 
plumosa, Artis, sp., is the only form I have hitherto met with in the Middle Coal 
Measures, where it would appear Dactylotheca dentata is either absent or very rare. 

It is perhaps better to leave the relationship of Dactylotheca dentata and Dactylotheca 
plumosa to each other an open question till the subject has been more fully investigated.* 

Alethopteris, Sternberg. 
Alethopteris lonchitica, Schloth., sp. 

Pecopteris lonchitica, Brongt., Hist. d. veget. foss., p. 275, pi. lxxxiv. figs. 1-7; pi. cxxviii. 
Alethopteris lonchitica, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 225, pi. xxxi. 
Filicites lonchiticus, Schloth., Flora d. Vorwelt, p. 55, pi. xi. fig. 22. 

Alethopteris lonchitica, Schloth., sp., var. decurrens, Artis, sp. 

Alethopteris decurrens, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 221, pi, xxxiv. figs. 2, 3 ; 

pi. xxxv. fig. lj pi. xxxvi. figs. 3, 4. 
Filicites decurrens, Artis, Antedil. Phyt., pi. xxi. 

Pecopteris Mantelli, Brongt., Hist. d. veget. foss., p. 278, pi. ixxxiii. figs. 3, 4. 
Pecopteris Mantelli, Lindley & Hutton, vol. ii. pi. cxlv. 
Pecopteris heterophylla, Lindley & Hutton, Foss. Flora, vol. i. pi. xxxviii. 

Remarks. — Zeiller treats Filicites decurrens, Artis, as a distinct species, but I am 
much more inclined to regard it as a variety of Alethopteris lonchitica, Schloth., sp. 

Alethopteris Serlii, Brongt. 

Pecopteris Serlii, Brongt., Hist. d. veget. foss., p. 292, pi. Ixxxv. 

Rhacophyllum, Schimper. 
Rhacophyllum crispum, Gutbier, sp., forma lineare, Gutbier, sp. 

Fucoides crispus, Gutbier, Vers. d. Zwick. Schwarzk., p. 13, pi. i. fig. 11 ; pi. vi. fig. 18. 

Rhacophyllum Lactuca, Schimper, Traite d. paleont. veget., vol. i. p. 684; pi. xlvi. fig. 1 ; pi. xlvii. fig. 1 

(fig. 2 1) ; vol. iii. p. 524. 
Fucoides linearis, Gutbier, Vers. d. Zwick. Schwarzk., p. 13, pi. i. figs. 10, 12. 

Remarks. — The specimens agree with the Fucoides linearis, Gutbier, which is most 
probably only a variety of Fucoides crispus, Gutbier. 

The fossil figured by Mr Marrat as " Rootstock of a Fern " t is a fragment of 
RJiacophyllum, and what has been mistaken for " palese or chaffy scales " are stainings in 
the matrix, and not an organic substance. 

Megaphyton, Artis. 
Megaphyton frondosum, Artis. 

Megaphyton frondosum, Artis, Antedil. Phyt., pi. xx. 

* Since this was written, I believe I have evidence to show that " dentata " is most probably only a variety of 
Filicites plumosa, Artis.— April 1889. t Proc. Liverpool Geol. Soc, Session 13, 1871-72, pi. i. fig. 2, 1872. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 411 

LYCOPODIACE.E. 

Lepidodendron, Sternberg. 
Lepidodendron Sternbergii, Brongt. 

Lepidodendron Sternbergii, Brongt., Prodrome, p. 85. 

Lepidodendron dichotomum, Sternb. (in part), Vers., i. fasc. i. pp. 19 and 23, pis. i., ii. (excl. pi. iii.); Vers., 

ii. p. 177, pi. lxviii. fig. 1. 
Lepidodendron dichotomum, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, pi. lxvii. fig. 1. 

Lepidodendron aculeatum, Sternb. 

Lepidodendron aculeatum, Sternb., Vers., i. fasc. i. pp. 20 and 23, pi. vi. fig. 2; pi. viii. fig. 1b; fasc. ii. 

p. 25, pi. xiv. figs. 1-4 ; fasc. iv. p. 10. 
Lepidodendron aculeatum, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 435, pi. lxv. 

Lepidodendron Haidingeri, Ettingshausen. 

Lepidodendron Haidingeri, Ettings., Steinkf. v. Radnitz., p. 55, pis. xxii. and xxiii. 

Remarks. — The collection contains several specimens of this species, of which I had 
previously only seen a single example from Shale over " Hughes' Vein," Cwmburla 
Colliery, near Swansea, which was collected by Mr Lewis, Swansea. 

Lepidostrobus, Brongt. 
Lepidostrobus variabilis, L. & H. 

Lepidostrobus variabilis, L. & H., Fossil Flora, vol. i. pis. x. and xi. 

(?) Lepidostrobus Olryi, Zeiller. 

Lepidostrobus Olryi, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 502, pi. lxvii. fig. 1. 

Lepidostrobus Geinitzii, Schimper. 

Lepidostrobus Geinitzii, Schimper, Traite d. paleont. veget., vol. ii. p. 62 (excl. syn. L. comosus, L. & H.). 
Lepidostrobus Geinitzii, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 501, pi. lxxxvi. fig. 2. 
Lepidostrobus variabilis, Geinitz (not L. & H.), Vers. d. Steinkf. in Sachsen, pi. ii. figs. 1, 3, 4. 
Lepidostrobus variabilis, Roehl (not L. & H.) (in part), Foss. Flora d. Steink. Form. Westph., pi. vii. 

%• 2. 
Lepidostrobus variabilis, Feistmantel (not L. & H.) (in part), Vers. d. bbhm. Kohlenab., pis. xliii. 

and xliv. 

Lepidophloios, Sternberg. 
Lepidophloios carinatus, Weiss. 

Lepidophloios carinatus, Weiss, Foss. Flora d. jiingst. Stk, u. d. Rothl., p. 155. 
Lepidophloios carinatus, Kidston, Catal. Palaeoz. Plants, p. 172. 



412 MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 

Halonia, L. & H. 
Halonia regularis, L. & H. 

Halonia regularis, L. & H., Fossil Flora, vol. iii. pi. ccxxviii. 

Remarks. — Halonia regularis is merely the decorticated fruiting branch of a Lepi- 
dophloios. 

Lepidophyllum, Brongt. 
Lepidophyllum lanceolatum, Brongt. 

Lepidophyllum lanceolatum, L. & H., Fossil Flora, vol. i. pi. vii. figs. 3, 4. 

Bothrodendron, Lindley & Hutton. 
Bothrodendron minutifolium, Boulay, sp. 

Bothrodendron minutifolium, Zeiller, Bull. Soc. Geol. de France, 3 e ser. vol. xiv. p. 176, pi. ix. figs. 1, 2, 

1885. 
Bothrodendron minutifolium, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 491, pi. lxxiv. 

figs. 2-4, 1888. 
Bothrodendron minutifolium, Zeiller, Veget. foss. du terr. houill. de la France (Extrait du Tome IV. de 

Vexplantion de la carte geol. de la France), p. 117, 1879. 
Bothrodendron minutifolium, Kidston, Trans. Geol. Soc. Glasgow, vol. viii. p. 65. 
Rhytidodendron minutifolium, Boulay, Le terr. houil. du nord de la France et ses veget. foss., p. 39, pi. iii. 

figs. 1 and 1 bis, 1876. 
Rhytidodendron minutifolium, Renault, Cours d. botan. foss., Deuxieme Annee, p. 52, pi. xii. figs. 1, 2, 

1882. 
Lycopodium carbonaceum, O. Feistmantel, Vers. d. bbhm. Kohlenab., Abth. ii. p. 9, pi. i. figs. 1 and 2 

(named on plate Lycopodites lycopodioides), 1875. 
Lycopodites selaginoides, Roehl (?wt Sternb.) (in part), Foss. Flora d. SteinJc. Form. Westph., p. 144, 

pi. vii. fig. 3, 1869. 
Lycopodites carbonaceus, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 495, pi. lxxiv. fig. 1. 

Remarks. — In a previous paper,* while recording this species from Lancashire, I 
called attention to the error I made in uniting Bothrodendron with Sigillaria dis- 
cophora, Konig, sp.t 

At PL II. fig. 6, is given a figure of a plant which was originally named Lycopodium 
carbonaceum by Feistmantel. The same fossil has been figured more recently by 
Zeiller (loc. cit.) under the name of Lycopodites carbonaceus, 0. Feistm., sp. I have 
fully satisfied myself, from the examination of foliage branches of Bothrodendron 
minutifolium, which were associated with stems bearing the characteristic marks of the 
species, that the Lycopodium carbonaceum of 0. Feistmantel has been founded on 
foliage branches of Bothrodendron minutifolium. Although small branches of this 
Lycopod, bearing the foliage, are not uncommon, yet owing to the dense manner in 

* Geol. Soc. Glasgow, vol. iii. p. 65. 

t See Annals and Mag. Nat. Hist., vol.xvi. p. 251, 1885. Note. — Since this paper was written I have seen 
specimens of Bothrodendron punctatum, L. & H., as well as of Bothrodendron minutifolium, from the Lower Coal 
Measures, Kilmarnock, Ayrshire. 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 413 

which the leaves clothe the stem, it is seldom that one finds examples which show the 
form of the leaves clearly. In the specimen I figure they are well shown at the point 
marked with a + , and are enlarged five times at fig. 6a. The leaves are linear 
lanceolate, single nerved, and taper to a sharp point. I have little doubt, however, 
that their form will vary somewhat on different parts of the plant. 

Sigillaria, Brongt. 
Sigillaria tessellata, Brongt. 

Sigillaria tessellata, Brongt., Hist. d. veget. foss., p. 436, pi. clvi. fig. 1 (?); pi. clxii. figs. 1-4. 
Sigillaria tessellata, Zeiller, Flore foss. d. bassin liouill. d. Valenciennes, p. 561, pi. lxxxv. figs. 1-9 ; pi. 
lxxxvi. figs. 1-6. 

Sigillaria mamillaris, Brongt., var. abbreviata, Weiss. 

Sigillaria mamillaris, Brongt., var. abbreviata, Weiss, Foss. Flora d. jiingst. Stk. u. d. Rothl., p. 165', pi. 
xv. figs. 1, 2. 

Sigillaria Arzinensis, Corda. 

PL I. fig. 2. 

Sigillaria Arzinensis, Corda, Flora Protogma, p. 29, pi. lix. fig. 12 (figure inverted). 
Sigillaria Arzinensis, Goldenberg, Flora Sarcepont. foss., Heft ii. p. 44, pi. x. fig. 14. 
Sigillaria Arzinensis, Kimball, Flora from the Apalachian Coal Field, p. 16, pi. i. fig. 5. 
Sigillaria Arzinensis, Schimper, Traite d. paleont. veget, vol. ii. p. 93. 

Description. — Ribs arched, separated by straight furrows ; leaf-scars ovate, lateral 
angles distinct but not very prominent ; cicatricules slightly above the centre, the two 
lateral straight or slightly lunate, diverging, the central punctiform. On the ribs, 
extending from one leaf-scar to another, is a longitudinal band of very fine punctato- 
granulate markings, and immediately above the leaf-scars are fine transverse corrugations. 
Subepidermal surface longitudinally striate. 

Remarks. — The surface ornamentations of the ribs of this species are very delicate, 
and unless the specimens are in a fine state of preservation they would not likely be 
shown. I have indicated the appearance they present at PI. I. fig. 2a. Above the 
scars the ornamentation is more pronounced, and assumes the form of fine transverse lines 
which bend round the upper end of the leaf-scar. On account of imperfect preservation, 
I can easily understand that this species will more frequently be found with a smooth 
bark, and in this condition it would be difficult to distinguish it from Sigillaria ovalis, 
Lesqx., # which may, perhaps, be only a form of Sigillaria Arzinensis, Corda. 

Stigmaria, Brongt. 
Stigmaria ficoides, Sternb., sp. 

Stigmaria ficoides, Brongt., Class, d. veget. foss., p. 9, pi. i. fig. 7. 
Stigmaria ficoides, L. & H., Fossil Flora, vol. i. pis. xxxi.-xxxvi. 
Variolaria ficoides, Sternb., Vers., i. fasc. i. pp. 22 and 24, pi. xii. figs. 1-3. 

* Goal Flora, vol. ii. p. 495, pi. lxxi. figs. 7, 8. 



414 MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 



Stigmaria rimosa, Goldenberg. 

Stigmaria rimosa, Goldeub., Flora Saraip. foss., Heft iii. p. 15 pi. xii. figs. 3-6 (named Stigmaria 
abbreviata on plate). 

Cordaites, Unger. 
Cordaites principalis, Germar, sp. 

Cordaites principalis, Zeiller, Flore foss. d. bassin. houill. d. Valenciennes, p. 629, pi. xciii. fig. 3, pi. xciv. 

fig. 1. 
Flabellaria principalis, Germar, Vers. v. Wettin u. Lobejun, p. 55, pi. xxiii. 

Antholithus, Brongt. 
Antholithus, sp. 

Sternbergia, Artis. 
Sternbergia approximata, Brongt. 

Sternbergia approximata, Brongt., Prodrome, p. 137. 

Sternbergia approximata, L. & H., Fossil Flora, vol. iii. pis. cexxiv., ccxxv. 

Remarks. — For notes on Sternbergia, see remarks under Araucaroxylon and Stern- 
bergia in the Catal. of Palceoz. Plants in the British Museum, pp. 220, 221. 

Trigonocarpus, Brongt. 

Trigonocarpus Noeggerathi, Sternb., sp. 
PI. II. fig. 4. 

Palmacites Noeggerathi, Sternb., Vers., i. fasc. 4, p. xxxv. pi. lv. figs. 6, 7. 

Trigonocarpus Noeggerathi, Brongt., Prodrome, p. 137. 

Trigonocarpus Noeggerathi, Zeiller, Flore foss. d. bassin houill. d. Valenciennes, p. 649, pi. xciv. figs. 8-11 

(excl. ref. Lindley & Hutton). 
Trigonocarpus Noeggerathi, Kidston, Trans. Roy. Soc. Edin., vol. xxxiii. p. 403, pi. xxiii. fig. 3. 
Trigonocarpus Noeggerathi, Kidston, Catal. of Pala30z. Plants, p. 216. 
Palmacites dubius, Sternb., Vers., i. fasc. 4, p. xxxv. pi. lviii. figs. 3a, b, c, d. 

Remarks. — Trigonocarpus Noeggerathi has been confused with Trigonocarpus 
Parhinsoni, Brongt., by Lindley and Hutton in their Fossil Flora, vol. ii. pi. clii c . and 
vol. iii. pi. cxciii 8 ., and pi. ccxxii. figs. 2, 3, where they figure seeds of the latter under 
the name of the former, and their figures have unfortunately been given as a reference 
to Trigonocarpus Noeggerathi by several authors. 

Trigonocarpus Noeggerathi is longer than Trigonocarpus Parhinsoni, the largest 
forms of the latter being seldom equal in size to the smallest specimens of the former, 
though in their length both species vary considerably. I do not know that Trigono- 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 415 

carpus Noeggerathi is ever less than 2*5 cm., if so small ; whereas the longest specimen 
of Trigonocarpus Parhinsoni that I have met with is 27 cm. long, but it was somewhat 
compressed. Of course, the diameter of Trigonocarpus Parhinsoni is proportionally 
smaller. 

Trigonocarpi have three prominent and three very slightly raised ridges placed 
between the stronger ones, but it is only on well-preserved examples that the slightly 
raised ridges or keels are observable. 

Of the two figures I have given of Trigonocarpus Noeggerathi, that in the Trans. 
Roy. Soc. Edin., vol. xxxiii. part ii. pi. xxiii. fig. 3, represents the form named dubius 
by Sternberg, which is, however, not specifically distinct from his Trigonocarpus 
Noeggerathi. 

Trigonocarpus Parkinsoni, Brongt. 

Trigonocarpus Parkinsoni, Brongt., Prodrome, p. 137. 
Trigonocarpus Parhinsoni, Kidston, Catal. of Palceoz. Plants, p. 217. 

Pinnularia, Lindley & Hutton. 
Pinnularia capillacea, L. & H 

Pinnularia capillacea, L. & H., Fossil Elora, vol. ii. pi. cxi. 
Pinnularia capillacea, Kidston, Catal. of Palmoz. Plants, p. 58. 

Stem. 

"Stem of a Fern," Marrat, Proc. Liverpool Geol. Soc, Session 13, 1871-72, pi. xi. fig. 1, 1872. 

Remarks. — A figure of one of these fossils has been given by Mr Marrat, who 
regarded them as fern stems. They possibly may belong to that group of fossils, but 
their state of preservation does not admit of a satisfactory determination. The stems 
bear spirally arranged obovate scars, which are considerably elevated in their upper part, 
but slope downwards towards their base, where they merge gradually into the stem, 
which is longitudinally striated. 



[Index, 
vol. xxxv. part ii. (no. 10). 3 z 



410 



MR ROBERT KIDSTON ON THE FOSSIL PLANTS IN THE 



INDEX TO SPECIES. 











PAGE 










PAGE 


Alethopteris, ... . .410 


Neuropteris, .... . 408 


loncliitica, .... 






410 


dentata, .... 








408 


,, var. decurrens . 






410 


gigantea, 








408 


Serlii, 






410 


heterophylla, . 








408 


Antholithus, ..... 






414 


macrophylla, . 








408 


sp., ... 






414 


tenuifolia, 








408 


Bothrodendron, .... 






412 


Odontopteris, .... 








409 


minuti folium, .... 






412 


Britannica, 








409 


Catamites, ..... 






398 


Reichiana, 








409 


(Calamitina) approximates, 






400 


Pecopteris, .... 








409 


(Stylocalamites) Cistii, 






401 


Miltoni, .... 








409 


(Calamitina) varians, Sternb., 


var. 


incon 




Pinnularia, .... 








415 


sfafts, Weiss, 






398 


capillacea, 








415 


(Euculamites) ramosus, 






400 


Rhacophyllum, 








410 


(Stylocalamites) Suckowii, 






400 


crispum, var. lineare, 








410 


( ,, ) undulatus, 






401 


Sigillaria, 








413 


(Calamitina) varians, var. 






400 


Arzinensis, 








413 


Calamocladus, 






401 


mamillaris, var. abbreviat 


%, 






413 


equisetiformis, 








401 


tessellata, 








413 


grandis, . 








401 


Splienopliyllum, 








401 


lyeopodioides, . 








401 


cuneifolium, 








401 


Cordaites, 








414 


Sphenopteris, . 








403 


principalis, 








414 


coriacea, . 








406 


Dactylotheca, . 








409 


Footneri, 








406 


plumosa, 








409 


furcata, . 








407 


Halonia, 








412 


Marratii, 








403 


regularis, 








412 


mixta, 








. 405 


Lepidodendron, 








411 


multifida, 








407 


aculeatum, 








411 


oblusiloba, 








405 


Haidingeri, 








. 411 


Sauveurii, 








. 403 


Sternbergii, 








411 


spinosa, . 








407 


Lepidopldoios, 








. 411 


Sternbergii, 








408 


carinatus, 








. 411 


trifoliolata, 








403 


Lepidophyllum, 








412 


Sphyropteris, . 








402 


lanceolatum, . 








412 


obliqua, . 








. 402 


Lepidostrobus, 








. 411 


Sternbergia, . 








. 414 


Geinitzii, 








411 


approximate/, . 








414 


Olryi, 








411 


Stigmaria, 








. 413 


variabilis, 








. 411 


ficoides, . 








. 413 


Mariopteris, . 








409 


rimosa, . 








414 


muricata, 








. 409 


Trigonocarpus, 








. 414 


Megaphyton, . 








. 410 


Noeggerathi, . 








. 414 


frondosum, 








410 


Parlcinsoni, 
Zeilleria delicatula, 








. 415 
. 403 



RAVENHEAD COLLECTION IN THE BROWN FREE LIBRARY AND MUSEUM. 417 

EXPLANATION OF PLATES. 
Plate I. 

Fig. 1. Catamitina (Catamites) varians, Sternb., var. inconstans, Weiss, nat. size. 

la. Portion of the node showing leaf-scars x 5. 
Fig. 2. Sigillaria Arzinensis, Corda, nat. size. 

2a. Leaf-scars and portion of rib enlarged 3 times, to show the surface ornamentation. 
Fig. 3. Sphyropteris obliqua, Marrat, sp., nat. size (type specimen). 

3a. Pinnule x 5, showing the pinnule segmentation and three sporangial bands. 

3b. A sporangial band x 7-^, showing the sporangia placed on oblique elevated ridges. 

3c. A sporangium x 53, showing the terminal depression (pore?) and the cell structure 
of the walls. 

3d. Another sporangium x 53. 
Fig. 4. Sphenojiteris coriacea, Marrat, nat. size (type specimen). 

4a, 4&. Two pinnules x 6. 

Plate II. 

Fig. 1. Sphenopteris Marratii, Kidston, n. sp., nat. size. 

la, lb, lc. Portions enlarged 5 times, to show the segmentation of the pinnules. 
Fig. 2. Sphenopteris Marratii, Kidston, n. sp., nat. size. 

2a. Pinnule x 5. 
Fig. 3. Splienopteris Footncri, Marrat, nat. size (one of the types). 

3a. Pinnule x 7^, to show the segmentation and nervation. 

3b. Terminal portion of a pinna x 7^. 
Fig. 4. Trigonocarpus Noeggerathi, Sternb., sp., nat. size. 
Fig. 5. Neuroptcris dentata, Lesqx., nat. size. (The small stem underneath the pinnule has no 

connection with it.) 
Fig. 6. Bothrodendron mimotifolium, Boulay, sp., nat. size. Small blanches showing foliage. 

6a. Two leaves x 5, from branchlet marked + on fig. 6. 



Trans. Roy Soc. Edin r Vol. XXXV. 

M R Robert Kidston on the Fossil Plants in the Ravenhead Collection in the Liverpool Museum. PI. I. 




t>" Kidston del 



ig-1-CALAMITINA (CALAM ITES) VARIANS, Sternb. var. INCONSTANS, Weiss. Kg. 2. SIGILLARIA ARZINENSIS, Corda.V 
Fig 3.SPHYR0PTERIS B LI QU A , Marrat, sp. Fig. 4. SPH E NOPT ER! S COR I AC E A , Marrat. 



Trans. Roy. Soc.Edin 1 Vol. XXXV. 

M K Robert Kidston on the Fossil Plants in the Ravenhead Collection in the Liverpool Museum. PI. II. 




i?- 



5* 








KiJsiim Jel 




19 









^ 



j> 



c^ 






F Huth, Lift 1 Edin 1 



N0Frr p SPHENOPTERIS MARRAT,I > Kidstonn.s. Fig. 3, SPH ENOPTERIS FOOTNERI, Marrat. Fig.4.TRIG0N OCAR PU S 
bGERATHI.Sternb. sp. Fig. 5. NEUROPTERIS DENTATA, Lesqx. Fig. 6. BOTH RODENDRON M INUTIFOLI U Nl, Boulay 



sp. 






%*, 



( 419 ) 



XI. — On some Fossil Plants from Teilia Quarry, Gwaenysgor, near Prestatyn, Flint- 
shire. By Kobert Kidston, F.K.S.E., F.G.S. (With Two Plates.) 

(Read 17th December 1888.) 

Teilia Quarry, from which the fossils described in this paper have been collected, is 
situated about a quarter of a mile north-east of the village of Gwaenysgor, which is 
distant about five miles from Ehyl and one from Prestatyn. 

The geology of the district has been fully described by Mr G. H. Morton, F.G.S., # 
from whose paper the following notes are compiled : — 

A considerable part of North Flintshire is occupied by a great development of lime- 
stone, which is generally referred to in writings on the subject as "Carboniferous Lime- 
stone." It forms a tract of hilly ground, some of the higher points rising to a height of 
over 800 feet. These limestones lie immediately beneath the Cefn-y-Fedw sandstones, 
and occupy a hollow of considerable extent in Silurian strata. 

From distinctive lithological characters, the limestones can be divided into four groups, 
all of which can be well studied in the steep escarpment near Prestatyn in the north- 
west of the county. 

In the north of Flintshire, the following are the four subdivisions of the limestones 
that occur: — 



Name. 


Thickness in 
Feet. 


Locality where well exposed. 


I. Upper Black Limestone, 
II. Upper Grey Limestone, 

III. Middle White Limestone, 

IV. Lower Brown Limestone, 

Red Basement Beds, .... 


200 
500 
600 
400 


Prestatyn and Llanasa. 
Coed-yr-Esgob and Gwaenysgor. 
Graig-fawr and Dyserth. 
Moel Hiraddug and Marian-cwm. 



Teilia Quarry is situated in beds which form the base of the Upper Black Limestones, 
— the uppermost subdivision of the series, — and in the present paper it is only with this 
subdivision we have to deal. 

This subdivision receives its name from the prevalence of black, fine-grained, thin- 
bedded limestones, which in weathering assume a light brown colour. This limestone 
was formerly worked for making hydraulic cement. 

According to Mr Morton, the original deposit was a calcareous mud, and rapid 
decomposition may account for so few organic remains being found. 

All the fossil plants from Teilia Quarry have suffered more or less from decay before 

* " The Carboniferous Limestone of the North of Flintshire," Proceedings Liverpool Geol. Soc, Sessions 1885-86, 
vol. v. pp. 175-197. See also Mem. Geol. Survey — " Geol. of Coasts adjoining Rhyl, Abergele, and Colwyn," (Explanation 
to Quarter Sheet 79 N.W.), by A. Strahan (with Notes by R. H. Tiddeman), 1885. 

VOL. XXXV. PART II. (NO. 11). 4 A 



420 MR ROBERT KIDSTON ON SOME FOSSIL PLANTS FROM 

fossilization intervened, and had probably spent some time on their water journey before 
finally becoming embedded, and, as further suggested by Mr Morton, were perhaps 
enveloped in a soft calcareous mud, in which for a period the ordinary processes of 
decay were carried on. Along with the plants, and on the same slabs, marine shells are 
frequent. 

Immediately overlying the plant-bearing strata is a bed with marine molluscan 
remains.* The whole series is essentially marine. 

From the remarks just made, it may be inferred that the plant remains do not 
generally occur in a good state of preservation, and with the majority of the specimens 
this is unfortunately the case. There are, however, many that can be easily identified, 
and a few more fortunate individuals which are in a very fair state of preservation. 
Some of these latter are here figured and described. 

As far as I can learn, the first collection of fossil plants from Teilia Quarry was made 
by the late Mr J. B. Shone, Chester, and it is now deposited in the Museum of the 
Society of Natural Science, Chester. Through the kindness of Mr A. W. Lucas, one 
of the secretaries of the Society, I have been enabled to examine these specimens, t 

Another good collection has been made by Mr E. Bouverie Luxmoore, F.G.S., St. 
Asaph, the better part of which he has contributed to the British Museum (Nat. Hist.), 
South Kensington, and to the Woodwardian Museum, Cambridge. I am indebted to 
Mr Luxmoore for the privilege of examining the specimens still in his possession, to 
Dr Woodward for examining those in the British Museum, and to Mr A. C. Seward, 
through whose kind offices Professor M'Kenny Hughes gave me an opportunity of studying 
the specimens presented to the Woodwardian Museum. I am also under additional 
obligations to Mr Luxmoore, by whom many of the best specimens were collected, and to 
Mr G. H. Morton, for assistance in points connected with the geology of the district in 
which Teilia Quarry is situated. 

Synopsis of Species. 

Asterocalamites, Schimper. 
Asterocalamites scrobiculatus, Schloth., sp. 

Asterocalamites scrobiculatus, Zeiller, Veget. foss. d. terr. Iwuil., p. 17., pi. clix. fig. 2. 

Calamites scrobiculatus, Schloth., Petrefactenhunde, p. 402, pi. xx. fig. 4. 

Catamites radiatus, Brongt., Hist. d. veget. foss., p. 122, pi. xxvi. figs. 1-2. 

Archwocalamites radiatus, Stur., Culm Flora, Heft i. p. 2, pi. i. figs. 3-8, pis. ii., iii., iv., v. figs. 1-2 ; 

Heft ii. p. 180, pi. ii. figs. 1-6, pi. iii. figs. 1-2, pi. iv. fig. 1, pi. v. fig. 1. 
Bovnia radiata, Schimper, Traite d. paleont. veget., vol. i. p. 335 ; vol. iii. p. 454 (Syn. iu part). 

* Lists of the fossils associated with this and the other subdivisions of the limestone are given in Mr Morton's 
paper. 

t I believe it was from this collection of fossils that the list of plants which was communicated by a friend to Mr 
Morton, and given by him at p. 179 of his paper, was drawn up. I am sorry to say, however, that only one of the 
identifications there given appears to be correct. 



TEILIA QUARRY, GWAENYSGOR, NEAR PRESTATYN. 421 



As its fruit — 



Pothocites Grantoni, Paterson, Trans. Bot. Soc. Edin., vol. i. p. 45, pi. iii., 1841. 

Fruit of Bornia radiata (Pothocites Grantoni), Kidston, Ann. Mag. Nat. Hist, 1883, vol. xi. p. 297, pis. 
ix., x., xi. figs. 9, 10 ; pi. xii. figs. 13-17. 

Remarks. — Not infrequent. Several specimens have been collected. 



Adiantides, Schimper. 

(Adiantites, Auct.) 

Adiantides antiquus, Ett., sp. (Plate I. fig. l.) 

Adiantides antiquus, Stur., Culm Flora, Heft i. p. 66, pi. xvi. figs. 4-6 ; pi. xvii. figs 3, 4. 
Adiantum antiquum, Ettingshausen, Foss. Flora d. mahr.-schles. Dachschiefers, p. 22, fig. 7, plate. PL vii. 
fig. 1 (in Denks. d. Kaiser. AJcad. Wiss, vol. xxv. p. 98, pi. vii. fig. 1). 

Description. — Frond tripinnate or decompound, pinnae alternate ; (?) secondary pinnae 
lanceolate ; tertiary pinnae, on lower secondary pinnae, lanceolate, on upper secondary 
pinnae broadly lanceolate, or composed of from 2-4 radiating pinnules. Pinnules 
cuneate, apex slightly rounded or truncate, entire or split into two or more cuneate 
segments. Veins numerous, dichotomous, and radiating from the contracted base of 
the pinnule, which forms a short foot-stock. 

Remarks. — On the lower and larger tertiary pinnae the pinnules are alternate, but 
the upper tertiary pinnae are mostly reduced to single pinnules, or to 2-4 pinnules, which 
are often arranged in fan-shaped groups, the component pinnules of which spring from 
a common point or have a much shortened rachis. The bipinnate condition of the 
secondary pinnae is therefore lost towards the apex of the primary pinnae. 

The fronds of Adiantides antiquus must have attained to considerable size. The 
specimen figured shows what are probably the remains of two primary pinnae, neither of 
which is perfect, the more imperfect being only represented by a small fragment on the 
lower left-hand corner of the figure. 

The nervation is very feebly indicated in the Teilia specimen, and that only in a few 
of the pinnules. From other specimens the nervation is shown to be very fine, the 
numerous dichotomous veins radiating from the base of the pinnule. The main rachis 
of a specimen figured by Stue # is nearly -^ inch wide, — this indicates the large size 
attained by the fronds of this species. Adiantides antiquus is rare in Britain. 

* Loc. cit., pi. xvii. fig. 3. 



422 MR ROBERT KIDSTON ON SOME FOSSIL PLANTS FROM 

Rhacopteris, Schimper. 

Rhacopteris flabellata, Tate, sp. (PI. I. fig. 2; PI. II. figs. 4-7.) 

Rhacopteris flabellata, Kidston, Catal. Palceoz. Plants, p. 63, 1886. 

Sphenopteris flabellata, Tate, in Johnston's Botany of the Eastern Borders,* p. 308, woodcut, fig. 3, 1853. 
Noeggerathia, sp., Gomes, Flora foss. do terr. Carbon do Porto, Serra do Bussaco, &c, p. 32, pi. ii. figs. 
1, 2, 1865. 

Description. — Frond pinnate (?) ; pinnse alternate, rhomboidal, or subtriangular in 
outline, much divided into narrow, linear, acute, single veined segments ; segments of 
pinnule generally arranged in fascicles, and the pinnule united to the rachis by a short 
foot-stalk. Rachis moderately thin. 

Remarks. — Owing to the fragmentary and imperfect condition of the specimen on 
which Tate founded his Sphenopteris flabellata, there is considerable difficulty in dis- 
tinctly knowing what are the real characters of his species. 

The woodcut figure of the type given by Tate represents portion of a pinnule, 
" petiolate, flabellate, deeply cut into three radiating segments ; segments cut more or less 
deeply into linear obtuse lobes, the lowermost having seven, the middle four, and the 
uppermost eight lobes ; veins radiating, one broad vein passing into each lobe." -J- 

From an examination of the woodcut, I am inclined to think that none of the seg- 
ments of the pinnule figured by Tate shows its true apex, but that all are embedded 
in their upper part or broken over. If I am correct in supposing this, there has only 
been a portion of a pinnule shown, and on this single specimen the species has been 
founded. 

In February 1886, when in the neighbourhood of Alnwick, I paid a visit to the 
Mechanics' Institute, where I believe Mr Tate's geological collection has found a 
resting place. My time at Alnwick was limited, but with as careful a search as 
circumstances would permit, we failed to find the type of Sphenopteris flabellata. 
Whether it is permanently lost, or only temporarily mislaid, is not at present known. 
Under these circumstances all that can be done is to try to interpret Tate's figure as 
best one can. It is questionable if, under such conditions, it would not be better to let 
such an imperfectly defined species sink into oblivion, but, on the other hand, one likes 
to give all honour to those who have been pioneers in a difficult branch of natural history, 
and to take every means for the preservation of their specific names, and I believe, 
notwithstanding the difficulty in identifying his species, that I am warranted in applying 
Tate's name of " flabellata " to the specimens now under consideration. 

For some years I have known of certain specimens which I thought were referable to 
Tate's Sphenopteris flabellata, the most perfect example being in the collection of the 
British Museum. This example is included in my Catalogue of Palaeozoic Plants as 
Rhacopteris flabellata, Tate, sp.J 

* Nat. Hist, of the Eastern Borders, London, 1853, vol. i. This is the only volume that was published, 
t Tate, loc. cit., p. 308. % Page 63. 



TEILIA QUARRY, GWAENYSGOR, NEAR PRESTATYN. 423 

I have also a small specimen collected by the late Mr C. W. Peach from the Oil 
Shales, West Hermand, West Calder, Midlothian (Calciferous Sandstone Series). 

While visiting the British Museum in 1887, Mr R. Bullen Newton, F.G.S., kindly 
showed me the collection of fossil plants which had lately been presented to the 
Geological Department of the British Museum by Mr E. B. Luxmooee, F.G.S., and 
among them were several specimens of the plant I believe to be Tate's Sphenopteris 
Jlabellata. These specimens, along with those to which reference has already been 
made, afford, I think, sufficient data for a satisfactory identification of all these fossils 
with Sphenopteris Jlabellata, Tate. This species must, however, be now placed in the 
genus Rhacopteris, Schimper,* of which the type is Rhacopteris (Asplenites) elegans, 
Ettingshausen. Rhacopteris was formed to include that group of Sphenopterids with 
rhomboidal, more or less divided or slit pinnules, which are attached to the rachis in a 
plane nearly vertical to the growth of the fern. 

Rhacopteris Jlabellata appears to have possessed a pinnate frond, the rachis of which 
may perhaps have divided into two equal parts like many other palaeozoic species, but 
this has not yet been observed. The pinnules consist of from 3 to over 20 linear, acute, 
single-nerved segments, the number of segments varying according to the position of 
the pinnule on the frond — those at the apex containing much fewer segments than those 
at the base. 

Fig. 2 is a reproduction of Tate's original figure, introduced here for comparison with 
those now given. 

Fig. 4 is a drawing of the most perfect example I have seen, and is from the 
Calciferous Sandstone Series, Burdiehouse, near Edinburgh. The uppermost pinnules 
only contain three segments, those towards the base eight or nine. 

Figs. 5-7 are from Gwaenysgor, near Ehyl. In fig. 7 the segments are closer and not 
so much expanded as in the other examples figured, having more of a matted appearance 
than is generally seen. 

The small specimen given at fig. 5 illustrates well the fasciculate arrangement of the 
pinnule segments. In fig. 6, which has suffered from maceration, the segments are 
longer than in any of the other Gwaenysgor examples. 

It is to be observed that in none of the specimens figured, nor in any that have come 
under my notice, are the pinnules so large as that given by Tate in the original wood- 
cut. This, however, may arise from his example originating from a larger individual, or 
from its having held a lower position on the frond than any of the specimens with which 
I have met. 

The plant described as Noeggerathia by Gomes t is probably referable to Rhacopteris 
Jlabellata. 

Schimper unites Gomes' plant with Rhacopteris elegans, Ett., sp.,| but as far as one 

* TraiU d. paUont. ve'ge't, vol. i. p. 481, 1869. 

t Flore fossile du terr. Carbon, des environs du Porto, Serra Bussaco, &c, p. 32, pi. ii. figs. 1, 2, Lisbon, 1865. 

+ Traite* d. paleont. ve'ge't, vol. i. p. 482. 



424 MR ROBERT KIDSTON ON SOME FOSSIL PLANTS FROM 

can learn from a study of Gomes' figures it appears more probably to belong to 
Rhacopteris flabellata, Tate, sp. 

Rhacopte?is flabellata differs from Rhacopteris Geikiei, Kidston,* in the former 
having unilateral pinnules with narrower segments, which do not spring from a distinct 
midrib as in Rhacopteris Geikiei. 

Rhacopteris insequilatera, Goppert, sp. 

Rhacopteris incequilatera, Feistmantel, Palceontograpliica Sripp., iii., Lief. iii. Heft ii. p. 74, pis. ii. fig. 3, 

iii. figs. 1, 2, iv. figs. 1, 2 ; Heft. iv. p. 145, pis. i. (xix.) figs. 3, 4, ii. (xx.) figs. 1, 3, iii. 

(xxi.) iv. (xxii.) figs. 2, 3, v. (xxiii.) figs. 4, 5, vi. (xxiv.) fig. 2 (1), 1878. 
Cyclopteris incequilatera, Gbpp., Foss. Flora d. Silur. Devon, u. unteren Kohlenf, p. 72, pi. xxxvii. 

figs. 6, 7, 1860. 
Adiantites Lindseaiformis, Bunbury, Geol. Survey of Ot. Brit. ; The Geology of the Neighbourhood of 

Edinburgh, pp. 144 and 151, fig. 26, 1861. 
Rhacopteris fiabellifera, Stur., Culm Flora, Heft i. p. 76, pi. vi. fig. 10. 

Remarks. — Several specimens are contained in the collections from Teilia Quarry, but 
all are more or less fragmentary. Goppekt's name for this species, being older than that 
given to it by Bunbury, must be adopted for this fern. I am, however, sorry to see the 
old and well-known name of Lindsewformis disappear from our lists. 

Archseopteris, Dawson. 

(?) Archseopteris, sp. 

Remarks. — There are one or two indistinct impressions in the collection, which appear 
to be referable to this genus. 



& v 



Sphenopteris, Brongniart. 

Sphenopteris subgeniculata, Stur., sp. (Plate II. fig. 10.) 

Diplothmema subgeniculatum, Stur., Culm Flora, Heft. ii. p. 135, pi. xii. figs. 8, 9, 10. 

Description. — Frond much divided, ultimate pinnae alternate ; pinnules alternate, 
and in the lower part of the pinna? composed of numerous simple or dichotomous, narrow, 
linear segments, into each of which extends a vein. Rachis slightly geniculate. 

Remarks. — This species is closely related to Sphenopteris geniculata, Germar and 
Kaulfuss,t but is smaller and more delicate in all its parts. 

Sphenopteris Teiliana, Kidston, n. sp. (Plate I. fig. 3.) 

Description. — Frond bipinnate (?) ; main rachis divided into two equal parts by an acute- 
angled dichotomy ; pinnae oblong, alternate ; pinnules alternate, fan-shaped, and composed 

* Trans. Roy. Soc. Edin., vol. xxx. p. 535, pi. xxx. fig. 5, pi. xxxi. fig. 9. 

+ Merkw. Pflanzenabdr. Verlmndl. d. k. Leop. Carol. Akad. d. Naturf., vol. xv. part ii., 1831, p. 224, pi. lxv. fig. 2. 



TEILIA QUARRY, GWAENYSGOR, NEAR PRESTATYN. 425 

of from 1-2 bifid cuneate segments. On the main rachis, below the bifurcation, are placed 
distant pinnules of the ordinary type. Nervation not shown. 

Remarks. — Fig. 3 shows the best preserved example, which is drawn natural size. It 
only shows one of the two arms of the dichotomy of the main rachis ; those specimens 
which showed the bifurcation of the axis were not in other respects so well preserved as 
the figured example. 

This species belongs to that group of ferns whose frond bifurcates into two linear seg- 
ments, a group which appears to be more common in Lower than in Upper Carboniferous 
rocks. 

The pinnules at the apex of the pinnae are simple or bifid. In those composed of four 
segments, the pinnule first forms a dichotomy, each half of which again dichotomises, 
the ultimate segments being slightly cuneate. Probably these were single-veined, but 
none of the specimens showed the nervation. 

The species is not infrequent at Teilia, but seldom at all well preserved. 

Sphenopteris pachyrrachis, Goppert. 

Sphenopteris pachyrrachis, Goppert, Foss. Flora d. Ubergangs., p. 143, pi. xiii. fig. 3, 1852. 
Sphenopteris pachyrrachis, Goppert, Flora d. Silur. Devon, u. unteren Kohlenf., p. 61, 1860. 
Sphenopteris pachyrrachis, var. stenophylla, Goppert, Foss. Flora, d. Ubergangs, p. 143, pi. xiii. figs. 4, 5. 
Sphenopteris pachyrrachis, Goppert, Flora d. Silur. Devon, u. unteren Kohlenf., p. 61. 
Archceopteris pachyrrachis, Stur., Culm Flora, Heft i. p. 64, pi. viii. figs. 8, 9. 

Remarks. — Both the typical form and the variety stenophylla occur at Teilia Quarry. 
The form named stenophylla does not really deserve varietal rank, as the differences 
between it and the typical plant are merely due to different positions on the frond. 

In my Catalogue of Palaeozoic Plants # I placed certain species under Sphenopteridium 
dissectum, in which was included Sphenopteris pachyrrachis. 

I still fail to see how many of these species are to be distinguished, but as I hope at 
an early date to have the opportunity of examining a larger series of these ferns than I 
have hitherto been able to study, the Teilia specimens are placed in the meantime under 
Sphenopteris pachyrrachis, with which they are certainly identical. 

If the figures of Sphenopteris crassa, L. and H., as already figured by me,t be compared 
with those of Sphenopteris pachyrrachis as given by Goppert, the close relationship of 
these two (?) species is very obvious. The whole matter of the affinities of these ferns to 
each other requires to be carefully gone into. 

(?) Sphenopteris Schlehani, Stur., sp. 

Calymmotheca Schlehani, Stur., Culm Flora, Heft. ii. p. 280, pi. xi. figs. 2-4. 

Remarks. — On the Teilia specimens only the outline of the pinnules is seen, and 
that but here and there. In general form the fern agrees closely with Sphenopteris 

* Page 61. 

t Proc. Roy. Phys. Soc. Edin., vol. vii. pi. v. Sphenopteris crassa, L. and H., was describe and figured in 1835. 



426 MR ROBERT KIDSTON ON SOME FOSSIL PLANTS FROM 

Schlehani, Stur., sp., but until more perfect specimens are discovered, the occurrence of 
this species must remain doubtful. 



Sphenopteris, sp. 

Remarks. — The various collections from Teilia contain a number of specimens of 
Sphenopteris. They are, however, too imperfectly preserved to admit of specific 
identification, though it is evident that they belong to several distinct species. 



(?) Fructification of Fern. (PL II. figs. 8, 9). 

Remarks. — Two small specimens of these fossils are shown on the Plate, natural size. 
In all there were five examples of these curious fossils from Teilia. One of them shows 
portion of a frond 4^ inches long, without exhibiting either extremity. The rachis is 
3 '5 mm. broad, and gives off alternate pinnae. Unfortunately this example has suffered 
much from imperfect preservation. 

The form of the (?) sporangia is seen in the two fossils figured. Fig. 9 shows a 
primary pinna (judging from the specimen to which reference has already been made). 
The (?) sporangia are here somewhat displaced, but the fossil gives some idea of their 
attachment to the pinna. In fig. 8 their attachment to the pinna is better shown. 

The (?) sporangia are curved-subclavate, their basal extremity being narrowed into 
a pedicel, their outer surface appears to be striated, but this appearance may arise from 
partial decay. They vary in length, as can be observed in the figures. 

Owing to the imperfect preservation of these interesting fossils, it is quite impossible 
to ascertain their minute structure with sufficient accuracy to warrant a generic or 
specific name being applied to them. 



Lepidophloios, sp. 

Remarks. — Among the specimens presented to the Woodwardian Museum by Mr 
Luxmoore, are two small fragmentary impressions of a Lepidophloios, but their imperfect 
condition does not admit of a specific identification. 



(?) Cordaites, sp. 

Remarks. — In Mr Luxmoore's collection is a fragment of a fossil, with parallel 
nervation, which is probably referable to this genus. The specimen is, however, too 
imperfect for a satisfactory determination. 



TEILIA QUARRY, GWAENYSGOR, NEAR PRESTATYN. 



427 



General Remarks. 

The following Table shows the distribution of the Teilia Quarry species in the 
Lower Carboniferous rocks of Scotland, with the horizon in which they occur : — 



Asterocalamites scrobiculatus, Schl., sp., . 
Adiantides antiquus, Ett., sp., . 
Rhacopteris flabellata, Tate, sp., 

„ incequilatera, Gopp., sp., 

(?) Archceopteris, sp., 
Sphenopteris subgeniculata, Stur., sp., 
Teiliana, n. sp., . 
pachyrrachis, Gopp., . 

„ forma stenophylla, Gopp., 
(?) „ Schlehani, Stur., sp., 

S P 

(?) Fructification of Fern, 

Lepidophloios, sp., ..... 

(?) Cordaites, sp., . 



Teilia Quarry. 



, Scotland. 



Carboniferous Lime- 
stone Series. 



(Genus). 
(Genus). 



Calciferous Sand- 
stone Series. 



'(Genus). 

(Genus). 
(Genus). 



Mr A. J. Jukes-Brown, F.G.S.,* suggests that the Upper Black Limestones in 
which Teilia Quarry is situated, may probably represent the Yoredale Group ( = 
Carboniferous Limestone Series of Scotland), but, from reasons to be given presently, 
I am rather inclined to think that these Flintshire beds belong; to a lower horizon than 
the Yoredales. 

If the species which have been specifically identified from Teilia Quarry be com- 
pared with the fossil flora of Scottish Lower Carboniferous rocks, it is seen that seven of 
the Teilia plants are known to occur in the Calciferous Sandstone Series of Scotland 
and only two in the Carboniferous Limestone Series of Scotland, which latter series is 
supposed to embrace the equivalents of the Yoredale Group. Although then admitting 
that a number of Lower Carboniferous plants are common to the Carboniferous Lime- 
stone and Calciferous Sandstone Series of Scotland, yet so many of the species that 
occur at Teilia have only been discovered hitherto in the Calciferous Sandstones that I 
believe the Teilia beds are more probably of Calciferous Sandstone than of Carboniferous 
Limestone age. Little or no assistance is given in deciding this point by the Molluscan 
remains, as some of them extend from the base to the top of the Carboniferous Forma- 
tion, whereas the Lower Carboniferous plants have a limited and definite distribution, 
none extending into the Upper Carboniferous, with perhaps one doubtful exception. 



* Students' Handbook of Historical Geology, p. 191, 1886. 
VOL. XXXV. PART II. (NO. 11). 



4 B 



428 MR ROBERT KIDSTON ON SOME FOSSIL PLANTS FROM TEILIA QUARRY. 

EXPLANATION OF THE PLATES. 

Plate I. 
Fig. 1. Adiantides antiquus, Ett., sp. 

Specimen in the Collection of the British Museum. Collected by Mr E. B. Luxmoore. Locality. — 
Teilia Quarry. 
Fig. 2. Rliacopteris flabellata, Tate, sp. 

Copy of the original figure in Johnson's Natural History of the Eastern Borders, p. 308, 1853. 
Locality. — Budle. 
Fig. 3. Spheywpteris Teiliana, Kidston, n. sp. 

In the Museum of the Society of Natural Science, Chester. Collected by the late Mr J. B. Shone. 
Locality. — Teilia Quarry. 

Plate II. 
Fig. 4. Rliacopteris flabellata, Tate, sp. 

Specimen in the Collection of the British Museum. Locality. — Burdiehouse, near Edinburgh. 
Horizon. — Calciferous Sandstone Series. 
Figs. 5-7. Rliacopteris flabellata, Tate, sp. 

In the Collection of the British Museum. Collected by Mr E. B. Luxmoore. Loccdity. — Teilia Quarry. 
Figs. 8-9. (?) Fructification of Fern. 

In the Museum of the Society of Natural Science, Chester. Collected by the late Mr J. B. Shone. 
Locality. — Teilia Quarry. 
Fig. 10. Sphenopteris subgeniculata, Stur., sp. 

In the Collection of the British Museum. Collected by Mr E. B. Luxmoore. Locality. — Teilia 
Quarry. 

All the figures are natural size. 



Trans. Roy. S . Edi b 

Kidston on Fossil Plants from Teilia Quarry. p LATF i. 




H-Kid B ton,de^ 



MfFarlanc kErstyne.Littf^Edtri 1 ' 



Fig.l. ADIANTIDES ANTIQUUS, Etb. sp. Fig 2. RHACOPTERIS FLABELLATA, Tate. sp. 

Fig. 3. SPHENOPTERIS TeILIANA, Kidsbon. n. sp. 



Trans. Roy Soc. Edm r , Vol XXX.V 
Kidston on Fossil Plants from Teilia Quarry. — Plate ii. 









10. 




R -Kidston, delV 



MTarlane &Ers"kine, Lith c f Ea> t 



Figs.-4-Z. RHACOPTERIS FLAB ELLATA, Tate. sp. Figs. 8"9. FRUCTIFICATION OF FERN.(?) 
Fig. 10. SPHENOPTERIS SUBGENICULATA, Star. sp. 



( 429 ) 



XII. — On the Behaviour of the Hydrates and Carbonates of the Alkali-Metals, and of 
Barium, at High Temperatures, and on the Properties of Lithia and, the 
Atomic Weight of Lithium. By Professor W. Dittmar. 

(Read 17th June 1889.) 

The fragmentary nature of our knowledge of the behaviour of the more strongly 
basilous hydrates and carbonates at high temperatures is owing chiefly to the absence of 
a suitable material for the necessary crucibles. Unfortunately there is no metal which 
combines the infusibility of platinum with the chemical inertness of gold, in opposition to 
fiery-fluid caustic alkalies. But the corrosive action of these on platinum, as I showed 
some years ago,* is a function only of the peroxides formed from them by the action of 
atmospheric oxygen, and, consequently, can easily be prevented by operating in an 
atmosphere of hydrogen or nitrogen. 

This observation forms the basis of the experimental methods used in the present 
research. Before commencing my report, however, I should wish to give myself the 
pleasure of acknowledging that almost all the laboratory work involved was done by my 
private assistant, Mr Eobert Anderson, and of expressing my indebtedness to him for 
the excellent manner in which he carried out my directions. 

Lithia. 

Having, in the course of my earlier research,t found that carbonate of lithia, if kept 
at a bright red heat in an atmosphere of hydrogen, loses carbonic acid very largely, I 
some time ago repeated the experiment, with the view primarily of seeing whether it is 
possible in this manner to obtain the pure oxide, Li 2 0, of which the handbooks of chemistry 
tell us so very little, and, in the case of success, utilising it as a starting-point for the 
preparation of the hydrate, of the properties of which we are almost equally ignorant. 
The experiment, as I may state at once, proved a success. 

The material which I started with this time was a supply of lithium carbonicum 
purissimum, from Trommsdorff of Erfurt, which, according to an analysis by Mr Robson, 
had the following composition : — 

Lithia, Li 2 0, ..... 
Soda, Na 2 0, ..... 
Potash, K 2 0, ..... 
Magnesia, MgO, .... 



Carbonic acid, C0 2 , . 
Sulphuric acid, S0 3 , 



VOL. XXXV. PART II. (NO. 12). 



40-14 : 


: 30-04 


= 


1-3362 


0-03 : 


: 62 


= 


0-0005 


0-07 


: 94 


= 


0-00074 


012 : 


:40 




0-0030 




1-34044 


59-32 


:44 


= 


1-3482 


0-09 


: 80 




o-oon 


99-77 


1-3493 


m. Industry for 1884, 


page 303. 


t Ibid. 








4 C 



430 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

The preparation, certainly, is not chemically pure ; yet I accepted it as sufficiently 
pure for my purposes.* 

As the preparation is very bulky, I began by fusing successive instalments in a 
platinum crucible (in hydrogen) to produce some 26 grammes of an anhydrous and 
partly causticised compact substance. The crucible used for this, and in the subsequent 
heatings, was the same as I had employed in my previous research, — an ordinarily shaped 
platinum crucible of about 30 c.c. capacity, provided with a well-fitting overlapping lid, 
bearing two autogynically soldered-in narrow platinum tubes, one for letting in the 
hydrogen, the other for the escape of the volatile products. The inlet-tube projects into 
the crucible to the extent of about 1 centimetre, the other commences at the lid. 

The fused preparation was found, in two analyses, to contain 42*19 and 42*12 
per cent, of C0. 2 , instead of the 59*44 demanded by Li 2 C0 3 , if Li = 7'02; = 16. 
15*41 grm. of this carbonate were placed in the crucible and heated repeatedly in a con- 
tinuous current of purified dry hydrogen, for, in all, thirty-seven hours. At first a gas 
blowpipe was used, but the constant working of the bellows proving tiresome, a simple 
powerful " Bunsen " was substituted and employed for the greater part of the process, the 
gas blowpipe being resumed only at the later stages. The total loss of weight amounted to 
8*562 grm., including a not inconsiderable quantity of volatilised lithia, which comes off 
more especially in the earlier stages of the process. The residual substance (6*849 grm.) 
was a very hard stone-like mass, which, when broken up into fragments and again heated 
in hydrogen, only sintered slightly, but did not fuse. It exhibited a slightly bluish tinge, 
owing probably to the presence in it of a trace of sulphide. For the analysis of the 
product a known weight of it was decomposed in a Classen apparatus (i.e., under an 
inverted condenser) with dilute hydrochloric acid, and the small quantity of C0 2 which 
came off, after having passed a chloride of calcium tube, collected in a tared soda-lime 
tube and weighed. 

The chloride of lithium solution produced was evaporated with sulphuric acid to dryness, 
the residue ignited, and these operations repeated until the sulphate was constant in 
weight and its solution neutral to litmus ; it was then weighed as Li 2 S0 4 . Three such 
analyses gave the following results : — 

(i) (2) (3) 

C0 2 . . . . 0-55 0*26 0-60 

Li 2 6 .... 99-60 99-57 not done. 

Adopted Numbers. 

C0 2 0-60 

Li 2 99-57 



100-17 



* In the research referred to I took special care to eliminate from my preparation every trace of potassium, 
rubidium, and caesium. Yet, when fused in platinum at a red heat in the presence of air, it attacked the metal badly, 
v. ith formation of a dark-coloured platinite. This is contrary to a statement of Troost's, who says (Victionnaire de 
Clrimie, vol. ii. part i. page 228), that " lithia" attacks platinum only if contaminated with rubidium or coesium. 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 431 

corresponding to the formula Li 2 + 0*004 Li 2 C0 3 , and showing that the preparation was 
a fair approximation to pure monoxide of lithium. 

In regard to the properties of this oxide, I have nothing to add to what I have stated 
incidentally, except that it acts only sluggishly on water with apparently little evolution 
of heat. An exact determination of the heat- evolution by means of a Bunsen calorimeter, 
and experiments on the action of the oxide on alcohol, are contemplated. 

Hydrate of Lithia. 

This, according to my original programme, was to have been produced from the oxide 
by means of water ; but the preparation of the latter having proved so very wasteful of 
time, I fell back upon the old method of decomposing the sulphate with baryta- 
water. 

160 grm. of Erfurt lithium carbonicum purissimum were dissolved in water, with 
addition of a slight excess of sulphuric acid ; the dissolved carbonic acid was expelled by 
heating, the surplus acid neutralised by addition of more carbonate, and the liquid 
filtered. On the other hand, an almost saturated baryta-water was prepared from pure 
crystals. 

Duplicate analyses of the two solutions showed that 1 grm. of sulphate of baryta 
was obtained (by addition of sulphuric acid and chloride of barium respectively) from 
25*109 grm. of the baryta-water, and from 3*8983 grm. of the lithia solution. Hence 
1 grm. of the lithium solution should demand 6*4412 grm. of the baryta- water. But 
a preliminary synthetical trial, made with 10 grm. of the lithia solution, showed that 
64*6 grm. of the baryta water were just, but barely, sufficient to bring down the whole 
of the S0 3 . Hence, in the actual preparation, the reagents were mixed in this latter 
proportion (in the cold), and allowed to stand in well-stoppered bottles until the 
precipitate of sulphate of baryta had settled completely. The supernatant liquors were 
then syphoned off, the precipitates mixed with boiled-out water, again allowed to settle, 
and the clear washings united with the first decantates, which, of course, had meanwhile 
been preserved in well-stoppered bottles. 

The united leys (which amounted to about 15 litres, and, by calculation, should have 
contained 58*77 grm. of Li 2 0) were boiled down as quickly as possible in a nickelled iron 
basin (as made in Iserlohn for culinary purposes) to about 1 litre, care being taken to 
keep the basin almost wholly covered with its lid during the evaporation, so as to leave 
only a small outlet for the steam. The concentrated liquor was then transferred to a 
bottle provided with a glass stopper, and next tested for S0 3 and BaO, when a trace of 
the former was found to be present. A small quantity of baryta water therefore was 
added, and the minute precipitate of sulphate of baryta allowed to settle. The clear 
liquor was now found to contain a very minute trace of baryta, which however was 
allowed to remain. 

The clear liquor was syphoned off and boiled down in a platinum basin (which was 



43'2 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

covered almost completely with a piece of sheet silver) to about 400 c.c, and the whole 
then allowed to cool under a bell jar trapped with baryta- water. The liquid soon froze 
into one mass of crystals, which was re-fused with addition of a little boiled-out water 
and again allowed to stand, when a large crop of small crystals separated out and 
left a tangible quantity of mother-liquor. The mother-liquor was decanted off, the 
crystals washed twice with small quantities of boiled-out water, and next sucked dry 
as far as possible by means of a Bunsen pump, a perforated cone of vegetable parch- 
ment serving as a filter. The funnel was covered with another funnel of the same size, 
the seam made tight with a strip of black indiarubber, and the stem of the upper funnel 
made to communicate with a gas-holder containing carbonic-acid-free air. After the 
mother-liquor had been thus removed as far as possible, the crystals were spread out on 
a plate of unglazed " granite- ware," and kept under a bell jar over 60 per cent, caustic 
soda ley, with the view of removing only the mechanically adhering water. When the 
crystals had become dry, in the sense of no longer making a mark on filter paper when 
pressed against it, they were transferred to a glass-stoppered bottle as " Prepara- 
tion (1)." 

The mother-liquor was allowed to evaporate under a bell jar over dry caustic soda, 
which led to the formation of relatively large crystals ; these were preserved as 
"Preparation (2)." 

The two preparations conjointly were of course bound to include almost the whole 
of what the original carbonate of lithia and the baryta-water used contained of ordinary 
alkali. To determine the latter in the baryta-crystals, 5 - 3 grm. of these were dissolved 
in water, the baryta eliminated as carbonate by means of carbonic acid gas (the 
bicarbonate being decomposed by heating and addition of carbonate of ammonia), and 
the filtered liquid evaporated to dryness. The small residue left was taken up with 
water, the residual carbonate of baryta filtered off, the filtrate again evaporated to dryness, 
and the residue ignited and weighed. It amounted to only 2 milligrammes, corresponding 
to 0*280 grm. for the total weight of baryta-crystals used in the preparation. Assuming 
those 2 mgrm. of residue to have been carbonate of soda, the 280 mgrm. correspond to 
164 mgrm. of Na 2 0. The total base alkali imported by the sulphate of lithia, according 
to Mr Robson's analysis, amounted to 150 mgrm. Hence the 58*77 grammes of Li 2 
contained in the total caustic lithia-liquor were contaminated with 0*314 grm. of base 
alkali (R 2 0), or about ^injth 0I> their weight ; and of this part at least must be 
presumed to have passed into the mother-liquor. Only Preparation (1) was analysed, 
in this way : — A known weight of substance was decomposed in a Classen apparatus 
with dilute hydrochloric acid, for the direct determination of the small quantity of 
carbonic acid present; in the residual liquid, or in separate portions, the baryta and 
the baryta + lithia respectively were determined as sulphates. Found in 100 parts — 

Lithia, Li 2 0, .... 34*12 and 34-09 Mean = 3410 

Carbonic acid, C0 2 , . . . 0'888 „ 0*930 „ = 0*909 

Baryta, BaO, .... 006 „ 0*09 „ = 0075 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 



433 



Hence by the means — 




* Lithia as hydrate, ...... 


. 33-48 


Lithia as carbonate, ...... 


0-62 


Carbonic acid in the carbonate, .... 


0-91 


Baryta, ........ 


0-08 


* Water, by difference, ..... 


. 64-91 



3410 



100-00 
* Corresponding to the formula Li 2 + 3-2347 H 2 0, or LiOH+ 1-1173 H 2 or rather LiOH.H 2 + 0-1173 x H 2 of 
surplus water. 

2 - 66 7 grm. of the crystals were placed in the gas crucible, and heated in a current of 

hydrogen, first gently and then over the gas blowpipe. After four hours' heating the weight 

of the residue became constant at 0*901 grm. = 3378 per cent, of the weight of the original 

substance, which comes very near the 34'10 per cent, of Li 2 found in the analysis. To 

confirm this result (which surprised me, because at the time I shared the general 

impression amongst chemists that even hydrate of baryta, BaO.H 2 0, is stable at a red 

heat) weighed quantities of the residue were converted into neutral sulphate and the 

latter weighed. Kesults — 

(1) (2) 

Substance analysed, . .... 0418 0*294 grm. 

Sulphate obtained, ..... 1-531 L075 „ 

Percentage of Li 2 0, .... 99-96 99-80 

The residue undoubtedly was anhydrous lithia. 

Mr Anderson's experiments on the behaviour of lithia-crystals at lower temperatures 
may be passed over because this subject was at a later date worked out more completely 
by Mr Henderson, who, however, had to make fresh lithia-crystals for his purpose, Mi- 
Anderson's stock having been entirely used, up in the solubility experiments, which I 
shall report on presently. A crop of crystals, corresponding to Mr Anderson's No. 2, was 
operated upon. 

For the determination of the crystal water, a known weight of crystals contained in a 
platinum boat was placed in a combustion tube surrounded by a square air-bath made 
of asbestos pasteboard, and in this condition kept at the fixed-upon temperature while 
a current of pure and dry hydrogen was passing through the tube. In each of the two 
experiments which were made we began by maintaining a temperature of 100° until the 
weight of the residue was constant. The residue was then exposed to a succession of 
higher temperatures for the times stated, and after each such period of heating the weight 
of the residue noted. 

Experiment I. — Weight of crystals taken = 1*1720 grm. Constant weight at 100° = 

0*6584 grm. ; hence loss = 0*5136. The weights of the successive higher temperatures 

was as follows : — 

After 40 minutes at 120° 0-6590 

150° 0-6598 

170° 0-6604 

200° 0-6602 



434 



PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 



Experiment II. — Substance taken = 1*1965 ; loss at 100° = 0*52535 grm. Weights 

of successive residues : — 

60 minutes at 100° 0*67115 

15 „ 220° 0*6710 

15 „ 250° 0*67125 

15 „ 320° 0*67105 

For the determination of the carbonic acid and of the lithia, a known weight of 
crystals was decomposed by hydrochloric acid in a Classen apparatus, and the C0 2 
absorbed in soda-lime and weighed. The residual chloride was converted into neutral 
sulphate, and the latter weighed, to be calculated into Li 2 0. 

(2) (3) 

0*8809 0*9966 

8*25 mgrm. 9*25 mgrm. 

0*897 " 0*936 0*928 

1*1560 not done 1*2560 

34-527 ... 34*385 



Substance, . 
C0 2 found, . 
Percentage of C0 2 , 
Sulphate obtained, 
Percentage of Li 2 0, 



(1) 
0*9135 

8*2 mgrm. 



From the means we have 

Lithia, Li 2 0, . 
Carbonic acid, C0 2 , 
Water volatile at 100°, 
Other water and errors, 



Per cent. 
34-460 
0*920 
43-864 
20-756 



Multiples of Formula- Values. 

0*0186Li 2 CO 3 + (Li 2 + 1*023 H 2 0) + 2 '163 x H 2 
of water volatile at 100° C. 



Considering that the number 1*023 in the bracket is burdened with a number of 
observational errors, and that (by an error of judgment) all the components were not as 
far as would have been possible determined in the same quantity of substance, we may 
accept it as a sufficient approximation to unity, and conclude that lithia-crystals lose their 
crystal-water at 100° and that the residual LiOH remains undecomposed up to 320° C. 

It is rather surprising to see a substance like LiOH lose its crystal- water completely 
at so low a temperature as 100° C. Lithia obviously in this respect comes nearer to 
baryta than to potash or soda, with which it is habitually grouped together. A similar 
result was arrived at in the determination of the solubility in water at a series of tempera- 
tures, on which I will now proceed to report. 

Experiments at 19°4. — 10 c.c. of boiled-out water were placed in a glass-stoppered 
bottle of about 50 c.c. capacity, a presumably more than sufficient quantity of crystals was 
added, the stopper put on and the bottle kept immersed in a water-bath, the temperature 
of which was maintained at 19°*4 (the "20" of an ordinary thermometer) for three 
hours. During the first hour the bottle was frequently agitated ; during the other two 
hours the bottle was left undisturbed, to enable the undissolved part of the substance to 
settle. A convenient quantity of the clear supernatant liquor was then drawn off with a 
pipette, previously tared, along with a test-tube just wide enough to accommodate it, and 
weighed (pipette and contents) in the test-tube. The solution was then promptly trans- 
ferred to a Classen apparatus, the carbonic acid liberated, collected in soda-lime, and 



Percentage of 
Li 2 + 3H 2 0. 


Li 2 + 3H 2 per 
100 of Water. 


16-02 


19-07 


18-63 


22-89 


18-85 


23-23 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 435 

weighed. The residual solution served for the determination of the lithia as sulphate. 
Three such experiments were made. The records of the first were as follows : — Weight of 
solution taken =6*395. C0 2 in the same =6*0 mgrm. = 4*1 mgrm. of Li 2 O = 10*l of 
Li 2 C0 3 . Hence corrected weight of hy drate-solution = 6' S9 5 — 0"0101 = 6*3849. Sulphate 
of lithia obtained 1*354 grm. = 0*3695 of Li 2 0. Hence Li 2 present as hydrate = 0*3654 ; 
and percentage of Li 2 in the saturated solution = 0*3654 x 100-*-6-3849 = 5*724, corre- 
sponding to 16*016 of the hydrate Li 2 + 3H 2 and 83*984 grm. of water ; or 19*073 grm. 
of Li 2 + 3H 2 per 100 grm. of water. 

Experiments II. and III. were carried out similarly, except that the lithia solution left 
after expelling the C0 2 was fractionated with the balance, to obtain duplicate determina- 
tions for the lithia. In both cases the two values for the lithia agreed very well with 
each other. The final results were as follows : — 

In the Solution saturated at 19°*4. 

Experiment. Percentage of 

I. . . 5-72 

II. . . 6-66 

III. . . 6*74 

Excluding No I. as being probably infected with an unobserved error, we have as a 
mean of II. and III. — 

6*70 18-74 23-06 

Experiments at 0°. — In the first two experiments (I. and II.) portions of the solution 
prepared at 1 9° *4 were kept in an ice-bath for about twenty-four hours, and quantities of the 
clear liquors drawn off for the determination of the C0 2 and Li 2 0. In Experiments III. and 
IV. 10 c.c. of boiled-out water were placed in a glass-stoppered 80 c.c. bottle, and next kept 
in ice for half an hour, to make sure of everything being at 0° ; a sufficiency of previously 
cooled-down crystals was then added, the bottle stopped up, the stopper tied over with 
vegetable parchment, and the bottle immersed in an ice-bath provided with a lid, 
and entirely buried in a mass of ice contained in an outer vessel. After forty-five hours' 
standing the ice in the inner bath was still in its original condition. A quantity of the 
clear liquor was drawn off with a cooled-down dry pipette, weighed and analysed as 
usual. The results were as follows : — 

In the Solution saturated at 0°. 

Experiment. Percentage of 

I. . . 5-77 

II. . . 5-61 

III. . . . 5-85 

IV. . . 6*32 



Percentage of 
Li 2 + 3H 2 0. 


Li 2 + 3H 2 per 
100 of Water. 


16*16 


19-28 


15-69 


18-61 


16-36 


19-56 


17-69 


21-49 



436 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

Thinking that No. IV. was affected with an unobserved error, I adopted the mean of 
I., II., and III., which is : — 

5-74 1607 19-15 

Experiments at 100°. — These were carried out as follows : — 4*6 grm. of lithia-crystals 
and 10 c.c. of boiled-out water were placed in a test-tube, which was then suspended in a 
steam bath, to establish and maintain a temperature of 100°. A " Thorpe stirrer" within 
the test-tube served to mix up the contents. The test-tube was provided with a per- 
forated cork, and the stem of the stirrer passed through a short bit of indiarubber 
tubing slipped over a short piece of glass tube accommodated in the perforation of the 
cork, so that the stirring could be done while the test-tube was closed. During the first 
hour the contents were being constantly agitated ; the stirrer was then withdrawn, and 
the contents allowed to clear up for five hours, at 100°. A small quantity of the 
clear liquor was now withdrawn by means of a hot pipette, the pipette with its contents 
quickly transferred to a test-tube containing some water, and the whole weighed. By 
subtracting the conjoint weight of pipette, added water and test-tube, we obtained the 
weight of the sample of solution taken out, which was analysed as usual. In the first 
experiment two samples were drawn — one after five hours' standing, another forty-five 
minutes later (Determinations I. and II.). In the second experiment a similar method was 
followed ; only, after withdrawing a sample of the clear liquor for Determination III., an 
additional small quantity of crystals, and a corresponding quantity of water were added 
to the residue, the whole heated for five hours and then a sample of clear solution taken 
out for Determination IV. 

When the sulphates of lithia were dissolved in water, a little silica was found to re- 
main in each case. This was filtered off and weighed, to be allowed for in the calculation. 
The silica amounted to about ^tnrth of the sulphate of lithia ; as it no doubt came out of 
the glass, the glass-alkali which accompanied it ought to have been determined, but I 
could not see my way to doing this satisfactorily, and therefore simply neglected it. 
The results were as follows : — 

In Solution saturated at 100°. 



Experiment. 


Percentage of 
Li 2 0. 


Percentage of 
Li 2 + 2H 2 0. 


Li 2 + 3H,0 

100 of Wat 


I. 


932 


26-09 


35-29 


II. 


9-39 


26-26 


35-62 


III. 


9-26 


25-92 


3499 


IV. 


9-34 


2614 


35-38 



Mean . . 9-33 2610 35"32 

Experiments at 50°. — In these the saturated solutions were prepared pretty much in 
the same way as in the 19° '4 experiments, viz., by shutting up the crystals with an 
insufficiency of boiled-out water in a close bottle, which was then immersed in a water- 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 437 

bath kept at 50°. The digestion was continued for six hours ; during the first two 
with frequent agitation. The withdrawing of the samples was effected exactly as in the 
case of the experiments at 100°. As in these, silica was found to have passed into the 
lithia sulphate ; it was allowed for as before. Results : — 

In Solution saturated at 50°. 

x- • . Percentage of Percentage of Li,0 + 3H,0 per 

Experiment. ^ Li 2 0+3H 2 0. 100 of water! 

I. . . 7-25 20-30 25-46 

II. 7-24 2026 25-40 

III. , 7-21 20-16 25-26 

IV. . . 727 20-33 25-52 



Mean, . . 7"24 20-26 25-41 

Summary of Results. — 100 parts of solution saturated at "t°" contains "y" parts 
of Li 2 as hydrate; hence, in it, 100 parts of water are combined with "S" parts 
the compound of Li0 2 + 3H 2 0. 

m . Values of y. Mean of the r ■., 

Temperature. Minimum. Maximum. Adopted Values. S from Mean y. 

0° . . 5-61 6-32 5-74 19-15 

19°-4 . . 5-72 6-74 670 23-06 

50°-0 . . 7-21 7-27 7-24 2541 

100°-0 . . 9-26 9-39 9-33 35-32 

When I entered the t's as abscissae and the mean y's as ordinates in a system of 
rectangular coordinates, I found that the four points suggested a curve, of which I found 
it difficult to admit that it represented the true relation, because it would have demanded 
a parabolic function of the third degree to translate it into an equation. I besides 
remembered, that the experiments at 0° had hardly had justice done to them in the sense 
of sufficient agitation of the mixture of crystals and water. Neglecting the y for 0°, the 
other three were found to fall in very well with eq. y = 6 '6375 — 0'002825£ + 0'0003f. 

The constant term is almost identical with the one rejected value for 0°, which 
strongly suggests that this may be the best value of the set after all. Of course the only 
way of deciding this question was to repeat the determinations at 0° with greater care. 
We accordingly did so, and, to make sure of everything, we repeated also the determina- 
tions at 19° "4. Only, as the carbonic acid in the solutions prepared at these temperatures 
amounts to very little, we omitted its determination this time, and only allowed for it by 
calculation, assuming that the weight of Li 2 present in a given solution prepared at 0° 
and 19° *4 respectively, per unit- weight of total Li 2 0, is the same as the corresponding- 
average for the determinations previously made at 0° and 19° '4 respectively. The 
experiments at 0° were conducted in the following manner : — 

I. 3*32 grm. of crystals, and 10 c.c. of boiled-out water were heated up to, and for an 

VOL. XXXV. PART II. (NO. 12). 4 D 



438 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

hour kept at, 50° with occasional agitation. The stirrer was now removed, the tube closed, 
and kept in a double ice-bath (vide supra) for twenty hours. Two successive portions of 
the clear liquid were drawn off with a pipette, weighed and analysed for Li 2 by evapora- 
tion with sulphuric acid, &c. The resulting percentages of Li 2 were, A = 6*713 
and B = G'687. 

II. 1*56 grm. of crystals and 5 c.c. of boiled-out water, both previously cooled in ice, 
were mixed in a test-tube standing in an ice-bath, and agitated constantly for forty-five 
minutes. The stirrer was then removed, the tube closed and kept within the double ice- 
bath for twenty hours. Two successive samples were then drawn off, and their lithia was 
determined as sulphate. The resulting percentages were C = 6700 and D = 6708. In 
the case of C there was a slight loss; the mean of the other three numbers is 6703. 
By applying the correction for carbonic acid, I obtained for the corrected number the 
value 6-667. 

Tlie New Experiments at 19° - 4. — I. The residue of crystals and solution saturated at 
0°, left after withdrawal of sample " B," was mixed with an additional half-gramme of 
lithia-crystals, the whole heated up to 50° and kept at this temperature for half an hour 
with constant stirring. The tube was then transferred to a water-bath kept at 19° '4, its 
contents constantly stirred there for about two hours, and then left at rest in the same 
bath for six hours, when the liquid was found to have cleared up completely. Two 
successive samples, E and F, were then withdrawn, and their Li 2 determined as usual. 

II. The residue left after withdrawal of sample D at 0°, after having been supple- 
mented by addition of a little water and a corresponding (excessive) quantity of crystals, 
was kept at 19 0, 4 for eight hours; during the first two with continual stirring. Two 
samples, G and H, were then sucked out and analysed. Eesults : — 

E f G H 

y = 6799 6-802 6798 6798 

Mean = 6799 ; or 6738 after correction for the carbonic acid. 
The new values for y at 0° and 19°'4, together with the old ones at 50° and 100°, 
fall in well with the formula : — 

y = 6-6750 — 0-00346Z + 0*0003^, as seen from the following comparison : # — 

¥ort= 0° 19°-4 50°-0 100° 

y by calculation, . 6-675 6721 7-282 9-327 

y by experiment, . 6-667 6-738 7-240 9-330 

For the preliminary determination of the solubility of caustic lithia in alcohol, a 
quantity of the crystals was kept in contact with alcohol of 97 per cent, by weight for 
six hours, at a temperature of 15° established by means of a water-bath. During the 
first two hours the mixture was being agitated occasionally, during the last four it was 
allowed to clear up. In weighed portions of the clear liquor the lithia was determined as 

* According to this formula there should be a minimum of solubility at < = 5 0, 77, and at ll 0- 54 the solubility 
should be the same as at 0°. At 5°-77, by calculation, y = 6-665. It would take very exact work to see whether this 
minimum has any existence outside the formula, which latter of course does not pretend to formulate the actual law in 
the relation between t and y. 



HYDBATES AND CARBONATES OF THE ALKALI-METALS, ETC. 439 

sulphate. In two experiments (which happened to agree absolutely) the percentage of 
Li 2 in the saturated solution was found to be 0'59. 

Experiments on the action of alcohol on the anhydrous oxide are in preparation. 

Before passing on I may say that a series of attempts to produce a peroxide of 
lithium from its oxide or carbonate failed. That metallic lithium when kindled in 
oxygen burns into chiefly peroxide is, I believe, ascertained. 

Knowing that fused carbonate of lithia at a red heat and in the presence of air 
attacks platinum crucibles very badly, I fully expected that oxide of lithium would 
unite readily with oxygen also in the absence of platinum. 

Other Hydrates. 

I. Baryta. — When I had found to my surprise that hydrate of lithia, when heated 
in hydrogen, is so readily reduced to oxide, I at once caused Mr Hodge to see how 
baryta-crystals behave under the same circumstances, and, as a necessary preliminary, to 
prepare carbonate-free crystals by dissolving a quantity of Erfurt "Purissimum" in 
boiling water, filtering the solution into a flask, allowing to cool, collecting the crystals 
which separated out on cooling on a funnel, and sucking them dry in a current of 
carbonic-acid-free air. The crystals were then more fully dried under a bell jar over 
solid caustic soda. 

A small quantity of them was placed in the "hydrogen-crucible," and therein exposed, 
in a current of hydrogen, first to lower temperatures and lastly to the full heat of a gas 
blowpipe. In the first experiment, the heating, after establishment of the highest 
temperature, was continued for two hours. The residue was then analysed by treating 
a known weight with water and then converting it into sulphate by evaporation 
with sulphuric acid. 0726 grm. of substance gave 1'068 grm. of BaS0 4 = 0*70148 of 
BaO = 96*62 per cent, of BaO. In the second experiment about 10 grm. of crystals were 
used, and the heating continued, until the weight of the residue became constant, which 
took about four hours. The residue looked like anhydrous baryta, and 0*342 grm. of it, 
when dissolved in water, and precipitated with sulphuric acid, yielded 0*5225 grm. of 
BaS0 4 , corresponding to 100*35 per cent, of BaO. 

These results of Mr Hodge's were subsequently confirmed by Mr Anderson, who 
besides extended the inquiry to the carbonate. 

1. About 2 grm. of baryta-crystals were heated in hydrogen over (ultimately) the 
blowpipe for an hour, and the residue was weighed; the heating was then resumed 
and continued for other thirty minutes, which, however, reduced the weight by only 
1 mgrm. 0*928 grm. of the product, when dissolved in water and hydrochloric acid, 
in a Classen apparatus, gave off 1 mgrm. of C0 2 = 0*108 per cent. The solution gave 
1*4096 grm. of BaS0 4 = 99*76 per cent, of BaO. 

(In regard to Carbonate of Baryta see below under "Carbonates.") 

After these results with the hydrates of lithia and baryta I was quite prepared to 



440 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

find that even the caustic alkalies proper, if kept at a red heat in hydrogen, suffer at 
least partial dehydration, and caused Mr Hodge to try some experiments with caustic 
soda to test the presumption. These experiments, however, proved more difficult than I 
had expected. There would have been no use in trying to dehydrate caustic soda at a 
temperature below redness ; and at a red heat, besides volatilising pretty fast, it "creeps" 
to an exasperating extent, so that after half an hour's heating (of some 2 to 3 grm. of 
substance), the crucible, when opened, was found almost empty, and the little of a 
residue that there was had to be scraped it before it could have been analysed, which it 
was impossible to accomplish without the substance absorbing water from the atmosphere. 
Mr Hodge analysed one or two of his products, but I did not preserve his numbers, because 
they did not prove the presence of real Na 2 0, and, in the circumstances, could not be 
taken as proving the absence of this component in the real product. 

Not feeling inclined to lose any more time over the matter, I simply left it on one 
side; but I subsequently found that the creeping of the fused alkali can be effectually 
prevented by burying it in a sufficiency of spongy platinum. A circular piece of platinum 
foil is turned into the shape of a cornet, whose seam is made fast and approximately 
tight by welding. The cornet is filled to about half its height with spongy platinum, 
the alkali put on the top and then covered over with more spongy metal so as to fill the 
cornet. The charged cornet is placed in the gas crucible and heated in the desired 
atmosphere. At the end of the experiment it is taken out, while still warm, with a 
forceps, quickly transferred to a wide weighing-tube provided with a hollow ground-in 
stopper, and weighed. For the determination of the carbonic acid and base it is placed 
in the decomposition-flask as it is. As the spongy metal forms a coherent mass, it is 
easy, at the end, to lift out the cornet and spongy metal, dry, ignite, and weigh it. I 
never came to use this method in connection with the hydrates, but I subsequently 
applied it occasionally in my experiments on the carbonates of the alkalies, and it is 
for this reason that I have here described it. 

Experiments with Carbonates. 

As early as 1860, Theodor Scheerer,* in the course of an investigation on the 
behaviour of silica to the carbonates of potash and of soda at certain approximately con- 
stant temperatures situated above their fusing-points, inquired into the behaviour of the 
unmixed carbonates under the same circumstances, and as a general result found that 
both reagents, when kept in a state of fusion within a well-covered platinum crucible, 
lost weight, the more largely the higher the temperature. The source of heat in all 
cases was a Berzelius spirit-lamp, i.e., a spirit-lamp constructed on the Argand principle. 
To produce " Rothyluth" the lamp was fed with spirit of 66° to 70° Richter; a 
somewhat higher temperature, "Orangegluth," was established by using 80° to 81° spirit, 
and keeping the level of the spirit constant. For the production of " Gelbgluth" the 
same lamp (and the stronger spirit) was used in conjunction with the blast arrangement 

* Likbig's Annalen, vol. cxvi. page 129. 



HYDRATES AND CARBONATES OF THE ALKALI- METALS, ETC. 



441 



2. 


3. 


4-994 


- 52 per cent 


4989 


0-62 


4-953 


1-34 


4-984 


0-72 


4-994 


0-52 


4-987 


0-66 g „ 


4-951 


1-38 "" „ 


4-985 


0-70 


4-993 


0-54 



known as "Planner's Spinne." The general modus operandi was, to take a known weight 
of the respective carbonate (previously dehydrated at a dull red heat) and expose it 
successively to the stated standard temperatures, taking care at each step to continue 
the heating until the weight of the residue was constant. The platinum crucible was 
provided with a close-fitting concave lid, to avoid loss by creeping and volatilisation 
as far as thus possible. His experiments with carbonate of soda may be quoted in detail. 
Weight of dehydrated carbonate operated upon, 5 '020 grm. These, when exposed 
successively to the temperatures named, in column 1 of the following Table, assumed the 
constant weights given in column 2 ; column 3 gives the losses of weight suffered by 
the carbonate, in percentages of its original weight. 

1. 

1. Eothgluth, 

2. Orangegluth, 

3. Gelbgluth, 

4. Orangegluth, 

5. Eothgluth, 

6. Orangegluth, 

7. Gelbgluth, 

8. Orangegluth, 

9. Eothgluth, 

The most remarkable feature in these results is that each of the three temperatures 
corresponds to very nearly a constant weight characteristic of the respective temperature ; 
the additional loss of weight, which took place in passing from 1. to 3., was undone by 
re-establishing the " Rothgluth" of stage 1. According to Scheerer "there can be no doubt 
that the matter which left the carbonate of soda in the heat, and (the matter) which 
was taken up again when it passed from higher to lower temperatures, was carbonic acid. 
It cannot reasonably be presumed to have been water.* We must admit that the ratio 
of the equivalents of soda and of carbonic acid is a function of temperature. Taking the 
equivalent of the former as constant and = Na 2 0, that of the latter, at "Gelbgluth," is 266, 
instead of 275 as at the ordinary temperature (0= 100). In solid carbonate of soda, as it 
is at ordinary temperatures, 1000 atoms of soda are combined with 1000 atoms of C0 2 ; 
at "Gelbgluth" 1000 atoms of soda unite with 967 atoms of carbonic acid." 

With carbonate of potash he obtained similar results. The salt, previously dehydrated 
at a dull red heat, lost 0*25 per cent, of its weight at " Orangegluth" and further 0*50 
per cent, on subsequent exposure to "Gelbgluth" and these 0'5 per cent, were taken up 
again when " Orangegluth " was re-established. 

This is the substance of Scheerer's results as far as they are in contact with the 
present investigation. 

In my own experiments the general order of operations was as follows : — In a first 
series of experiments the respective carbonate was exposed for a shorter or longer time to 
a red heat in an atmosphere of pure hydrogen, which in all cases led to the production of 

* ? Vide infra. 



44:2 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

a residue containing caustic alkali. This effect must be presumed d priori to be owing to 
two causes, namely, firstly to the dissociating tendency of the heat, and secondly to the 
reducing action of the hydrogen on the C0 2 of the carbonate (Ex. Li 2 C0 3 4- H 2 = Li 2 + 
CO + H>0). Hence, to form an estimate of the share of the heat in the change, a second 
series of experiments was made, in which the salt was heated in an atmosphere of nitrogen. 
When I tried this for the first time with carbonate of soda, I was surprised to obtain 
a residue which contained its caustic alkali chiefly, if not entirely, as NaOH. Where did 
the hydrogen of the NaOH come from ? Graham's famous experiments on the occlusion 
of gases by hot platinum afford an easy answer. The hydrogen came out of the flame 
and through the walls of the red-hot crucible. The effect of this adventitious hydrogen 
could not have been minimised by establishing ever so rapid a current of nitrogen, 
because the substance in the crucible is in a state of fusion, and the hydrogen comes to it 
chiefly direct through the walls, and not by way of the atmosphere, of the crucible. Hence 
in the subsequent nitrogen-experiments the modus operandi was brought into the 
following form : — The carbonate was placed in a platinum boat, which was then shoved 
into a porcelain tube and within it heated in a muffle, matters being arranged so that 
the boat was at the hottest point of the muffle. The muffle was heated by means of gas 
in a Fletcher furnace. 

Carbonate of Baryta. 

1. Hydrogen-Experiments. — Carbonate of baryta, prepared from pure chloride by 
precipitation with carbonate of ammonia, and dehydrated exhaustively at a dull red heat, 
was placed in the gas crucible, and in it heated ; for the first two hours by means of a 
Bunsen, then over a gas-blowpipe. The weights of the successive residues was as 
follows : — ■ 

Original Substance, .... 3427 



After three hours' heating, 
After four „ „ 

After five „ „ 

After six „ „ 

After seven 



2-7952 

2-667 

2-665 

2-671 (weighing error ?) 

2-667 



0"9592 grm. of the product, by decomposition with dilute hydrochloric acid, gave 
1 mgrm. of C0 2 , or 0404 per cent, (direct method) ; the solution yielded 1*4462 grm. 
of BaS0 4 corresponding to 0'9498 of BaO or 99'021 per cent, of the weight of the sub- 
stance analysed. The deficit of 0"875 per cent, must be water, which, however, was 
probably taken up between the last heating and the weighing out of the sample for 
analysis, because hydrate of baryta, as I have shown, is reduced to BaO at a red heat. 
The product was practically free of carbonic acid, and this is the main point. 

Nitrogen-Experiment. — 2"2168 grm. of carbonate were heated in a current of nitrogen 
in a porcelain tube within a muffle as above explained, the heating, after attainment of 
the highest temperature, being continued for one hour. The product weighed only 
1 # 9 mgrm. less than the original carbonate, which is sufficient to show that carbonate of 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 443 

baryta is not subject to dissociation at the degree of red heat which prevailed in our 
muffle. To make, however, quite sure of this result, the carbonic acid in the product 
was determined by the direct method, and the baryta in the residual solution as sulphate. 
The 2-2149 grm. of product gave 0*4928 of C0 2 , and 2*6182 of BaS0 4 . Hence— 

Found. Calculated from BaC0 3 . 

Carbonic acid, . . . 22-249 22-312 

Baryta, . . . 77-635 77-688 



99-884 100-000 

Carhonate of Lithia. 

This carbonate, as we have seen, is completely reduced to anhydrous oxide, if kept at 
a red heat in hydrogen for a sufficient time * the last instalments of the carbonic acid, 
however, go away very slowly, which indicates the existence of a basic carbonate 
relatively difficultly reducible by hydrogen. The following two 

Nitrogen- Experiments were made with the view of seeing whether, or to what extent, 
a similar decomposition can be brought about by the action of heat alone. In both the 
carbonate was heated within a porcelain boat in a muffle as described above. 

In Experiment I, the heating, after attainment of the maximum temperature, was 
continued for one hour (?). # 

0'6024 grm. of the product gave 0*3303 of C0 2 and 0*9983 of Li 2 S0 4 corresponding 
to 54*83 per cent, and 45*22 per cent, respectively, or to the following proximate 
composition : — 

Carbonate of lithia, ..... 92'26 
Caustic lithia, Li 2 0, ..... 7 - 74 



100-00 

Experiment II. — Carbonate taken = 0*8486 grm. ; the maximum temperature 

having been reached, the heating was continued for five hours. The residue then 

weighed by 0*1360 grm. less than the original carbonate. After other five hours' heating 

the total loss was 0*2728, which indicated that the change had not reached its end yet. 

Being, however, anxious to know the state of matters at that stage, I caused Mr 

Anderson to stop the experiment and analyse the product. 0*5484 grm. gave 0*2268 of 

C0 2 and 1*1780 of Li 2 S0 4 , whence we had — 

Lithia, . . . 58-620 Carbonate of lithia, . . 69*592 

Carbonic acid, . . 41-357 Oxide of lithium, . . 30-385 



99-977 99-977 

corresponding to the formula Li 2 0-f 0*4817CO 2 , or approximately to Li 2 C0 3 + Li 2 0. 
This goes a certain way towards confirming my surmise as to the existence of a stable 
basic carbonate ; yet I am inclined to think that all the carbonic acid can be expelled by 
heating in nitrogen, provided only we heat long enough. 

* Note lost. 



444 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 



Experiments in Carbonic Acid. 

From our general experience regarding decompositions like the one under 
consideration,* we have a right to presume that the decomposition of carbonate 
of lithia at a red heat can be prevented by maintaining over the heated carbonate 
an atmosphere of carbonic acid of sufficient density, and from the fact that in 
the two (nitrogen) experiments quoted the carbonic acid came off only very slowly it 
appears probable that even at a bright red heat the requisite carbonic-acid density, and 
consequent carbonic-acid pressure, does not amount to much, perhaps to less than one 
atmosphere? This idea formed the basis of the following experiments: — 

Experiment I. — A quantity of carbonate of lithia was placed in the gas crucible and 
kept for an hour at a red heat, over an ordinary Bunsen lamp, while a current of dry 
carbonic acid was made to pass through the crucible. The product, which had the aspect 
of a dull-looking glassy fuse (a glassy structure or rather non-structure, is characteristic 
for fused normal, in opposition to ordinary fused, carbonates of soda and potash — vide 
infra), was tested in a roughly quantitative manner with chloride of barium, and found 
to contain rather less than 0"2 per cent, of caustic lithia, Li 2 0. 

I here refer to a method for the detection and determination of caustic in carbonates of alkali, 
which Mr Henderson worked out under my direction. A known weight of the carbonate is 
dissolved in boiled-out water in a stoppered bottle, a sufficiency of boiled-out solution of chloride of 
barium is then added, and the mixture allowed to stand in the absence of air, until the precipitate 
has settled as far as possible. An aliquot part of the clear liquor is now drawn off {not filtered off) 
and titrated with very dilute standard-solution of hydrochloric or sulphuric acid in the presence of 
phenol-phthalein as an indicator. I prefer to fractionate gravimetrically, because this does not tie 
one down to a pre-determined volume of mixture. By means of this method (which, by the way, 
does not pretend to be absolutely original) we had no difficulty in detecting considerably less than 
01 per cent, of caustic potash or soda in the respective carbonate ; and a solution of absolutely 
normal carbonate {vide infra) always gave a perfectly neutral decantate. But when we proceeded 
to test it quantitatively, with solutions prepared from known weights of normal carbonate, and 
known volumes of recently-prepared standard-solutions of caustic alkali, the general result, at least 
in all those cases in which the caustic alkali amounted to 1 per cent, or less, was that only some 
two-thirds of the RHO passed into the barytic solution. I accordingly put down the method as 
being exact only in a qualitative sense. But we subsequently came to use it frequently side by 
side with the gravimetric method (determination of the C0 2 and of the total B 2 0), and always found 
the two methods to give fairly agreeing results. Hence, in all probability, the caustic alkali 
solutions used in the synthetical trials, all precautions notwithstanding, contained part of their 
alkali as R 2 C0 3 . 

To give a more exact idea of the method as applied to carbonate of lithia, I may quote the 
analysis of a carbonate " Q " which will be referred to below as an important preparation. 
1-825 grm. of this substance was dissolved in about 300 c.c. of boiled-out water in a glass-stoppered 
bottle standing over a water-bath. After all the carbonate had dissolved, which took a long time, 
the solution was allowed to cool to almost ordinary temperature, and then mixed with recently 
boiled-out solution of 10 grm. of chloride of barium. The bottle was then left to itself for about 
twenty-four hours, when the precipitate had settled very well. The bottle was then weighed, when the 

* See my memoir on the "Oxides of Manganese" in the Proc. Roy. Soc. Edin., for 1864, also Chem. Soc. Jour., 
same year, p. 294 ; also Debray's " Experiments on Carbonate of Lime," Jahresb. for 1867, p. 85. 



HYDRATES AND CARBONATES OF THE ALKALI- METALS, ETC. 445 

weight of the total contents was found to amount to 46L5 grm. ; 212 - 3 grin, of the clear liquor, 
when titrated with a hydrochloric acid containing 0"02 x HC1 mgrm. per c.c, and dilute baryta- 
water as an auxiliary reagent, was found to require 1*78 c.c. of the acid for its exact neutralisation, 
corresponding to 3 - 86 c.c. for the whole, or to - 063 parts of free Li 2 per 100 of original substance. 

Experiment II. — Done in a similar manner to No. I. ; only the temperature was kept 
at the lowest point, enabling the carbonate to be completely fused. The product, this 
time, was analysed exactly with chloride of barium, and found to contain 0*018 mgrm. of 
free Li 2 in 320 mgrm., or 0*0057 per cent. 

Experiment III. — In the case of this experiment the heat was moderated so that the 
carbonate was near to, but below, its f using-point. The product, when tested with chloride 
of barium, was found to be perfectly neutral. This method subsequently was largely 
employed for preparing anhydrous and yet Li 2 0-free carbonate of lithia for other trials. 

Experiment IV. — A platinum boat charged with carbonate of lithia was placed in a 
combustion tube strengthened by means of a spiral of annealed brass foil wrapped round 
it, and in it heated by means of a combustion furnace to redness, while a current of 
carbonic acid was made to pass through the tube, and from it through a layer of mercury 
2 inches deep. The tube blew out shortly after a red heat had been fairly established, 
but the product was found to have been in a state of complete fusion, and therefore was 
analysed by means of chloride of barium. It contained only 0*081 mgrm. of free Li 2 
in 758 mgrm., or 0*01 percent. I have no doubt in my mind that the product really 
was neutral, and only became slightly caustic through the sudden removal of the carbonic 
acid atmosphere at a red heat. Or, in other words, that at the temperature which 
prevailed in the tube, and probably at any degree of " red heat," the dissociation-tension 
of carbonate of lithia, though decidedly greater, is only very little greater than 30 inches 
of mercury. But, to prove this, it was desirable to have a closer approximation to pure 
Li 2 C0 3 than our Erfurt "purissimum" afforded. I therefore caused Mr Anderson to 
purify some of this preparation by means of Troost's method, i.e., by dissolving it in 
water with the help of carbonic acid, and, from the filtered solution, reprecipitating the 
dissolved salt by removal of the loosely combined carbonic acid at a gentle heat. The 
crystalline crust thus obtained was washed with small instalments of water, dried at a 
gentle heat, and preserved for the following experiments. When it appeared desirable to 
operate upon an anhydrous preparation, a portion of the purified salt was placed in a 
platinum boat, and kept at a temperature somewhat below its fusing-point, within a 
combustion tube in a current of dry carbonic acid of ordinary pressure. 

Experiment V. — This was conducted pretty much like the preceding experiment, 
only with the difference that a porcelain tube was substituted for the glass one. The 
tube was heated in a combustion furnace, and after attainment of the highest temperature 
the heating continued for one hour. The depth of mercury through which the carbonic 
acid bubbled out was 2 inches, and the barometer stood at 29*78 inches ; hence the total 
pressure of the gas equalled 31 '78 inches. This pressure, at the end of the experiment, 
was maintained until the product had cooled down considerably below its fusing-point. 

VOL. XXXV. PART II. (NO. 12). 4 E 



44(5 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

The carbonate had fused into a white mass, which was smooth and glassy on its surface, 
but full of small holes on the other side where it had been in contact with the boat. 
0416 grm., when analysed by means of chloride of barium, was found to be neutral. 
0*8390 grm. gave 0*5000 of carbonic acid, equal to 59*594 per cent. The number 
calculated from the formula Li 2 C0 3 is 59*427, if Li be taken as =7*02, = 16. The 
analysis, as we see, gave about 0*16 per cent, more, although our preparation could not be 
presumed to be absolutely free from foreign bases, and any of these of course was bound 
to depress the percentage of carbonic acid. Admitting that the difference of 0*16 per 
cent, was not merely an observational error (need I say that I did not by any means 
feel sure on this point ?), the experiment would prove that Stas's number for Li (7*02) is 
a little too high, because there can be no reasonable doubt that Dumas and Stas's value 
44 for CO2 is practically free of error. Combining what I had found with the fact that 
both Mallet and Teoost, in 1857, by the analysis of the chloride with nitrate of silver 
found values for Li which lay decidedly below 7*00 (see Lothar Meyer and Seubert, 
Atomgeivichte der Elements, p. 225), I deemed it worth my while to prepare normal 
carbonate of lithia in the way described on a large scale, to analyse it with the requisite 
degree of precision, and to see what would come out. We accordingly repeated 
Experiment V. a sufficient number of times to obtain a sufficiency of material and then 
analysed the carbonate in the following fashion. To avoid repetition let me premise 
the statement that the same method was used for all the carbonic-acid determinations 

quoted below as "A," " B," "C" "Q," and considered in connection with the 

question of the constant " Li." 

A known weight of the carbonate (about 5 grm.) was decomposed with dilute 
sulphuric acid in a flask provided with an inverted condenser (Classen's method), which 
by its outlet communicated with (1) a large chloride of calcium tube for the absorption 
of the water; (2) a Liebig's bulb-apparatus, # charged with a solution of caustic potash in 
its own weight of water; (3) a medium-sized U-tube filled with granulated soda-lime, and 
(the greater part of the outlet-limb) with granulated chloride of calcium of the same kind 
as that used for (l) ; and (4) one or two smaller U-tubes charged with soda-lime and 
chloride of calcium like No. (3), to serve as " witnesses" of the complete absorption of the 
carbonic acid. In the first two analyses (quoted below as A and B) only one witness- 
tube was employed. It gained 0*5 and 1*2 mgrm. per gramme of substance analysed 
respectively. This is not much, but it is more than can be neglected in an atomic- weight 
determination ; I therefore, in the subsequent analyses, added a second witness-tube. 
This second tube gained weight, per gramme of carbonate analysed, to the extent of — 

0*1 mgrm. or less in nine cases, 
0*31 to 0*11 mgrm. in three cases, 
1*64 mgrm. in one case, and 

2*7 mgrm. in another case, namely, the case of analysis of " Q (3)" which was rejected on 
other grounds. 

* On the occasion of these analyses I invented an improved form of the LlEBlG bulbs, which is described and 
figured in the Hoc. Clitm Inch Jour, and also in the Chernilcer Zeiturtg, for 1888. 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 447 

The absorption apparatus were tared in two ways, namely, firstly, in the ordinary 
way by means of weights, and secondly, in the Regnault fashion, i.e., by means of 
similar apparatus of a slightly less weight, and of very nearly the same displacement. 
The tare potash-bulbs were charged with plain water, the tare U-tubes with granulated 
selenite. In the calculations I relied on the results of the Regnault method ; the 
other, indeed, was used only as a guarantee against weighing-blunders. I may say, in 
passing, that the two values for the weight of C0 2 found always agreed to within less 
than 1 mgrm. 

The apparatus having been put together, and the tightness of the joints made sure of, 
the carbonate was decomposed slowly by addition of dilute sulphuric acid, the Liebig 
bulbs being kept cool by occasional application of a water-bath ; when the carbonate was 
fully decomposed, the absorbed part of the gas was expelled by heating the contents of 
the flask to very near boiling, and passing a current of carbonic-acid-free air through the 
apparatus at the same time. This was continued until an approximate calculation, based 
upon the known volume of air used and the volume of the empty space of the apparatus, 
showed that the carbonic acid was sure to have all been driven into the absorption 
apparatus. These were then detached, immersed for a time alongside of the tare 
apparatus in a water-bath of the temperature of the balance, wiped dry, and weighed. 
Each apparatus and its tare were kept in the balance-case, with the ends closed, until 
their difference of weight was constant, the caps which closed the ends of the U-tubes 
and of the potash-bulb being removed in the actual weighing not having been included 
in the tare. The weighings here, as in most of the analyses quoted in this memoir, were 
made with a set of iridio-platinum weights from Messrs Johnson, Matthey, & Co., but 
adjusted by Mr Oertling. When the set came to me, I determined the errors of the 
several pieces by a series of comparisons and readjusted a few of the pieces which needed 
it. The balance used was that very fine " Hectogramme-balance " from Oertling, 
which I refer to in my article " Ueber die Pracisionswaage des Chemikers," in the 
Zeitschrift fur Instrumentenkunde for 1881, p. 113 et seqq.; the instrument is provided 
with the microscopic reading arrangement, which I contrived some years ago and described 
in the same Journal (for 1882, p. 63). 

As already stated, the first set of analyses was made with a carbonate, which, after 
dehydration in carbonic acid, was fused in an atmosphere of the same gas of about 
32 inches pressure. As a general result, the percentage of carbonic acid found was higher 
than that demanded by the formula, if Li = 7 '02 (Stas), and = 16. This confirmed my 
surmise regarding the true value of the constant, apart from the consideration that my 
carbonate, having been produced in an atmosphere of carbonic acid, may have contained 
physically absorbed, in addition to its combined, C0 2 . Assuming this to be so, the 
weight of C0 2 present per unit-weight of oxide of lithium should be governed by 
an equation like y = A + Bp, where p stands for the C0 2 -pressure under which the 
substance was fused and A for the chemically combined part of the carbonic acid. 
To find the values of the constants A and B, I prepared a few samples of the 



448 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

fused carbonate under pressures of about 45 and 60 inches, and analysed these like 
the rest. 

In all the experiments now to be reported on, the carbonate of lithia operated on was 
contained in a platinum boat standing in a porcelain tube, which was heated by means of 
a powerful gas-combustion furnace. The inlet and the outlet tube was provided with a 
glass stop-cock and inserted by means of a good indiarubber cork. When the pressure 
within exceeded that of the atmosphere without by not more than 2 inches of mercury, 
the corks were found to hold tight by friction ; but in the experiments at over-pressures 
of 1 5 inches or more they would surely have been blown out, had they not been held fast 
by a suitable arrangement. The contrivance which I used consisted of two perforated 
iron plates, united into a narrow parallelogram by two iron rods running parallel to the 
porcelain tube. When the apparatus is used, each of the plates lies flat against the 
surface of the respective cork, the inlet or outlet tube passing through the perforation. 
One of the plates is riveted to the rod-ends, the other is adjusted to, and held in, its 
position by means of thumb-screws. 

In the experiments at higher pressures the outlet tube dipped into a layer of mercury 
about 2 inches deep ; the entrance-tube communicated, more immediately, with one of two 
globular mercury-reservoirs united (in the Geisler pump fashion) by means of a long 
piece of stout indiarubber tubing. This reservoir, I will call it the "working bulb," 
communicated in its turn with the final outlet of the " Kipp's apparatus," which 
furnished the stream of dry carbonic acid, the connecting tube being provided with 
a stop-cock. The other reservoir, the " reservoir proper," sat on a ring of a tall retort- 
stand, so that it could be placed at greater or less heights. To enable the " Kipp " to give 
the necessary pressure for driving the gas through the said 2 inches, or occasionally 
through a greater depth of mercury, the acid in it is adjusted so that, supposing the gas- 
evolution to have been going on for a while and the outlet tube then to be closed, the 
acid driven into the upper bulb fills it almost completely ; and this bulb again, by means 
of a tube fixed in its neck and dipping into the acid, and a piece of indiarubber tubing- 
slipped over the end of the glass tube, communicates with a bottle having a tubulus near 
its bottom, through the latter. 

In a given experiment, the first step always was to raise the reservoir so as to fill 
the working bulb almost completely, and the second to fill the tube with carbonic acid 
of about 2 inches more than the existing atmospheric pressure, the reservoir being 
lowered occasionally and re-raised to make sure of all the air being expelled. While a 
continuous stream of carbonic acid passed through the apparatus, the tube was heated up 
gradually, so as to expel the moisture from the preparation at the lowest sufficient 
temperature, but at last brought up to the highest temperature which the combustion 
furnace would afford. When the carbonate could be presumed to be completely fused, 
and, under the circumstances, saturated with carbonic acid, the working bulb was, if 
still necessary, filled as far as possible with carbonic acid. The outlet-tube and the tube 
connecting the working bulb with the " Kipp " were then closed, and the reservoir 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 449 

raised so as to bring the carbonic acid within the porcelain tube up to the desired 
tension, which was maintained as far as practicable by occasional readjustment of the 
height of the reservoir. 

Counting from the time when the tube had come up to the full heat afforded by the 
furnace, the experiments at to 2 inches over-pressure were continued for about one hour; 
in the high-pressure experiments (at 15 or 30 inches over-pressure) the initial pressure (of 
atmosphere + 2 inches) was kept up for about one hour, and the subsequent high pressure 
for another half hour. At the end of an experiment, after the gas flames had been turned 
off, the high pressure was always maintained until the temperature of the tube had sunk 
so far that the carbonate inside could safely be assumed to have frozen. 

In now proceeding to give an account of the individual experiments, I will take 
these up in what I conceive to be the most convenient order for the reader; the reference 
marks give the order in which they were made. 

Experiments Q. — A quantity of carbonate of lithia was placed in the gas-crucible and 
in it kept for 1 hour at a temperature just sufficient to fuse it, but no more, while a 
current of carbonic acid of ordinary pressure was passing through the crucible. This 
operation was carried out three times ; the three products were powdered roughly, mixed 
together, and divided into four parts, of which three were weighed out at once along 
with their respective preparation tubes, to be analysed for carbonic acid. The fourth 
quantity was kept in reserve for emergencies. As one of the three analyses gave an 
incomprehensible result, the fourth was broken into for an additional determination of 
carbonic acid. The results were as follows : — 





(1) 


(2) 


(3) 


(4) 


Substance, 


5-3387 


5-6930 


5-9016 


2-3752 


Carbonic acid, 


3-1760 


3-3847 


34793 


1-4125 


Percentage of C0 2 , . 


59-490 


59-454 


58-955 


59-465 



Two portions of (4), each equal to about 1'8 grm., were analysed exactly with chloride 
of barium. Found in one sample 0*063 percent, of Li 2 ; in the other, 0*049 and 0*056. 
Omitting 0*049, which was obtained by a titration done on a very small scale, we have 
0'060 for the percentage of free Lithia. Of the four carbonic-acid determinations 
that in (3) was rejected as being in all probability infected with an unobserved grave 
error; the mean of the other three is 59*470; hence so much C0 2 is contained in 
100 — 0*060 = 99*94 parts of carbonate proper; hence the latter contains 59*506 
per cent, of carbonic acid. 

Experiment E. — A quantity of the carbonate was dried in a current of carbonic 
acid at 250°, at ordinary pressure. 4*9432 grm. gave 2*9500 grm. of C0 2 , or 59*678 
per cent. 

Experiments A, B, and C were made with a supply of fused carbonate produced 
under carbonic acid of 3 1 *8 inches pressure in instalments. 

The analyses gave the following results : — 



450 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 





A. 


B. 


C. 


Substance, . 


5-2502 


5-2520 


5-0723 


Carbonic acid, 


31315 


3-1265 


3-0259 


Percentage of C0 2 , 


59-645 


59-529 * 


59-655 



Experiment D. — Carbonate kept fused under carbonic acid of 30*24 inches pres- 
sure for one hour. The chloride of barium test proved the absence of caustic lithia. 
Carbonic acid not determined. 

Experiment F. — Substance prepared as in case D, only the outgoing gas was not 
made to bubble through mercury; the barometer stood at 29*2 inches. 67288 grm. 
gave 3*9559 grm. of carbonic acid, or 58*790 per cent. This number almost proves 
the presence of caustic lithia, but to make sure of this the chloride of barium test was 
applied (this time in only a roughly quantitative fashion); it showed that there was 
upwards of 0*6 per cent, of free lithia. 

Experiment G. — A quantity of carbonate was fused under carbonic acid of (Bar. + 
extra mercury =) 30*24 inches pressure for one hour. The experiment was made twice 
to obtain a sufficiency of material for an analysis on a large scale. 4'2220 grm. gave 
2*5198 grm. of carbonic acid, or 59*683 per cent. 

Experiment H. — Carbonate of lithia was fused under carbonic acid, first for 
one hour at a pressure of 30*1+2 = 32*1 inch, then for half an hour at a pressure 
of 60*1 inch. 1*2984 grm. gave 0*7809 grm. of carbonic acid, or 60*143 per cent. 

Experiment J. — Done in the same way as H ; actual pressure in the final stage = 59*94 
inches. 3*8206 grm. of product gave 2*2845 grm. of carbonic acid, or 59*794 per cent. 

Experiment K. — Again, by intention, a repetition of H. Final pressure = 59*96 
inches. 2*527 grm. of product gave 1*5057 grm. of carbonic acid, or 59*584 per cent. 

Experiment L. — Carbonate of lithia was heated in carbonic acid, first for one hour 
under 2 inches of extra pressure, then for half an hour under 15 inches: or about 45 
inches of total pressure. This product was not analysed, because it consisted, part of 
a solid glassy fuse, part of glassy particles, and part of a dull powder. In Experiments 
J and K also the product did not present itself as a compact fuse, but in the form of 
glassy particles, which looked as if they had been produced from what was originally a 
fuse by spontaneous disintegration on cooling. But my subsequent experience enables 
me to affirm that fused carbonate of lithia, even if produced under pressure, is not 
subject to such disintegration. The aspect of product L showed that the temperature 
which prevailed in the tube, in that experiment at least, was not sufficient to produce 
a homogeneous fuse, hence the apparatus was transferred to another room which afforded 
a better gas pressure, and the work continued there. For 

Experiments M, N, and P, the substance for analysis was prepared as in the case 
of H, except that the pressure in the final stage of the heating process was not 
necessarily the same. The exact values of these pressures, and the results of the 
analyses, are given in the following table : — 

* This analysis did not proceed quite regularly; yet I have no reason to suspect that the result is at all far out. 



M. 


N. 


P. 


59-95 


44-7 


59-73 inch. 


4-3870 


4-2880 


4-373 grm. 


2-6174 


2-5558 


2-6027 grm. 


59-663 


59-604 


59-518 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 451 

Pressure of C0 2 in tube, . 
Substance for analysis, 
C0 2 obtained, .... 
Percentage of C0 2 , . 

Experiment O. — Carbonate of lithium fused for one hour in a porcelain tube under 
carbonic acid of 29*5 + 2*2 = 31 '7 inches pressure. 4*8665 grm. of the product gave 
2*9005 grm. of carbonic acid, or 59*601 per cent. 

The products M, N, 0, and P were compact homogeneous glassy fuses. 

We will now proceed to discuss these results, and first of all notice that, in the 
porcelain tube experiments, only one, namely, Experiment F, which was made at 
29*2 inches of carbonic-acid pressure, gave a product containing free lithia ; in all the 
rest, made at pressures from 31*7 inches upwards, neutral (or perhaps quasi acid) products 
were obtained. This shows that at the temperature which prevailed in the porcelain 
tube in these experiments the dissociation-tension of carbonate of lithia lies between 
29 *2 and 31*7 inches of mercury. Another glance at the numbers shows that the carbonic 
acid taken up by the heated carbonate did not increase with the pressure of the carbonic- 
acid atmosphere under which it is prepared. The numbers, if taken as they stand, 
would rather point the other way. The mean of the percentages of carbonic acid found 
for products A, C, G- and O, for which the pressure varied from 317 to 30*2, is 59*646, 
while the mean of the high-pressure experiments M, N and P is 59*595. 

We will now proceed to view the analyses as so many determinations of the value 
Li 2 0, and begin with the case of Q, because these analyses afford an upper limit for the 
constant. A substance proved to contain free lithia gave in three well-agreeing analyses, 
(1), (2) and (4), 59*470 per cent, of carbonic acid; hence the weight of lithia which in 
it was associated with C0 2 = 44 parts of carbonic acid, amounts to 44 x 40*53-^59*470 = 
29*987 parts of lithia, and this, if the substance were normal carbonate, would be the 
value of Li 2 0, whence we should have Li = 6*993. But the substance analysed, 
according to our analysis by chloride of barium, contained 0*060 per cent, of free lithia. 
Taking this number to be correct, we have Li 2 0= 13*943, whence Li = 6*971. This 
latter number must of course be taken for what it is worth, but the value calculated 
from the uncorrected carbonic-acid determinations proves that Li must be less than 6*993. 

We will now pass to the analysis of product E, which ought to afford a lower limit 
for the true value Li, because the substance may have contained more, but cannot have 
contained less, than the proportion of carbonic acid corresponding to normal carbonate. 
From the percentage of carbonic acid found (59*678) we calculate Li 2 = 29*729 and 
Li = 6*865. But this calculation is based upon only one analysis. 

The following table summarises the results of all the porcelain tube experiments, 
excepting F, which of course is out of court. The second column gives the pressure of 
the carbonic-acid atmosphere under which the respective carbonate was produced. 



4.VJ 



PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 



Experiment. 


Pressure. 


. 


31-7 


A . 


31-8 


B . 


31-8 


C . 


31-8 


D . 


30-2 


G . 


30-2 


N . 


44-7 


P . 


59-7 


M . 


60-0 


H . 


601 


J . 


59-9 


K . 


60-0 



Percentage of C0 2 
59-601 
59-645 
59-529 
59-655 



Li 2 = 



29-824 
29-770 

(29 - 913) suspected exp. 
29-757 



Not determined, but substance proved neutral. 

59-683 29-723 

59-604 29-821 

59-517 29-928 

59-663 29-748 

60-143 (29159) 



59-794 
59-584 



29-586 
29-845 



In addition to B, I exclude Experiment H, because the percentage of carbonic acid 
found is far higher than that obtained in any other of the high-pressure experiments, and, 
besides, the analysis was made with an exceptionally small quantity of substance. The 
mean of the other (nine) numbers is 29 "7 78 ; from the deviations of the individual 
results from the mean, I calculate that the "probable" error of the mean is = ± 0*020. 
Hence Li = 6*889±0'010. If we allow only the low-pressure experiments (O, A, C, G) to 
vote, we have Li 2 = 29769 and Li = 6*8842±0*007 ; but we have no excuse for excluding 
the high-pressure experiments. Another question is whether we may not neglect Experi- 
ments J and P as being in all probability infected with abnormal errors. In their case, 
indeed, the deviations from the mean amount to about three times the probable error of a 
single determination, which is ± 0*06. If we do exclude these two results, the mean of 
the remaining seven becomes equal to 29 '7 8 4 for Li 2 and 6 '892 for Li with a probable 
error of ± 0*0136 for the former. I adopt this number 6*892 =b 0*007, or rather 6*89 = 
Li, as the net result of my work, the more unhesitatingly as it, after all, does not differ 
much from the general mean of the nine experiments. But, how does it agree with the 
results of previous observations by others ? Of these the following chiefly come into 
consideration : — * 

I. Mallet, in 1857, analysed chloride of lithium in two ways, namely — 
A. By determining the chlorine gravimetrically as chloride of silver ; results of two 
analyses — 



Li - 



Minimum. 
6-947 



Maximum. 
6-950 



Mean. 
6-949 



B. By titrating the chlorine with silver ; one analysis gave Li = 6*934. 

II. Troost, in 1857,t analysed carbonate of lithia dried, "sometimes in vacuo, 

* I quote from Lothar Meyer and Seubert, Die Atomgewichte der Elemente, Leipzig, 1883. With Lothar 
Meyer and Seubert, "Mean" means not the arithmetical mean of the several determinations, but the result as it, 
comes out if all the analyses are calculated as one. Thus, for instance, in the case of four analyses of chloride 
of lithium by nitrate of silver, the four quantities of chloride are added together, and compared with the sum 
of the four precipitates of chloride of silver. 

t Lothar Meyer and Seubert refer to a memoir published in 1857. I find, in the Zeitschrift fur Chemie for 
1862, an abstract of a memoir of Troost's on the same subject, in which the numbers quoted as immediate data of the 
analyses agree with those given by Lothar Meyer and Seubert. 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 453 

sometimes at 200°/' by igniting a known weight with powdered quartz and determining 
the loss of weight. He found, in two analyses — 

Minimum. Maximum. Mean. 

6-986 7-021 7-008 

III. Diehl, in 1857, analysed the carbonate, after drying it at 130°, by decomposing 
it with dilute sulphuric acid and ascertaining the loss of weight. Found in four 
analyses — 

Minimum. Maximum. Mean. 

7-013 7-037 7-027 

IV. Teoost, in 1857,* analysed the chloride by weighing the chlorine as AgCl. Two 
analyses gave — 

Minimum. Maximum. Mean. 

6-944 7-002 6-966 

V. Stas analysed the chloride three times by titration with silver, and found — 

Minimum. Maximum. Mean. 

7-023 7-028 7-026 

VI. Stas converted the chloride into the nitrate and determined the ratio of 
LiCl : LiN0 3 ; he found in three analyses — 

Minimum. Maximum. Mean. 

7-026 7-035 7-030 

In contrasting these several results with my own, Li = 6*892, I naturally begin with 
the results of Diehl's and Troost's analyses of the carbonate, and note that they agree 
with one another but differ from me, Diehl by about -f 0*135 units =19 times my 
probable error, and Troost by +0*115 = 16 times my probable error. But Diehl 
determined his carbonic acid by means of a method of insufficient precision, and Troost 
dried his carbonate " sometimes in vacuo, sometimes at 200°," and yet produces only two 
analyses ! Besides, it is not proved that carbonate of lithia loses all its water at ordinary 
temperatures in vacuo, nor is it proved that it retains all its carbonic acid in air at 200°. 
Supposing the negatives to hold, his value for Li must be too high. Hence his results, 
no more than Diehl's, can be said to be incompatible with mine. 

In now passing to Mallet, I must admit that his three analyses of the chloride agree 
fairly well, and yet the mean of the two " means " 6*942, is greater than my result by 0*05 
units = to seven times my probable error. But, at Mallet's time Dumas's excellent 
method for at the same time dehydrating and deoxygenating a metallic chloride (by heating 
it in HC1 gas) was not invented yet; his chloride was bound to contain oxychloride, 
and his analysis to give too high a value for Li. 

But, unfortunately for me, Stas's result is still higher than Mallet's, and his chloride 
cannot be presumed to have been contaminated with oxychloride, because, if his silver- 

* See footnote, page 452. 
VOL. XXXV. PART H. (NO. 12). 4 F 



454 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

analysis (as quoted under V.) had, on this account, given too high a result for Li, his 
determinations of the ratio LiCl : LiN0 3 would have done the reverse, while in reality 
the two results are almost identical. It would be ludicrous for me to try and minimise 
the strength of Stas's evidence against the correctness of my number, but, then, I cannot 
get over the fact, that my substance "Q," although proved to contain free Li 2 0, and, in 
all probability not absolutely free of foreign metals, in three well-agreeing analyses, gave 
percentages of carbonic acid corresponding to Li = 6*981, 7'004, and 6*997 respectively, 
without allowing for the free lithia, or for impurities. If my preparations had been 
absolutely free of magnesia, soda, &c, my value for Li would have become less than it 
did. If I have gone wrong, it can have been only through two causes : — My carbonate, 
having been produced in carbonic-acid atmospheres of from 30 to 60 inches tension, may 
have contained surplus C0 2 , perhaps in the form of a pyro-carbonate Li 2 C0 3 + C0 2 ; or 
my method for the determination of the carbonic acid may be infected with a constant 
positive error. To settle these doubts, I caused Mr Anderson to make a few experiments 
with perfectly pure carbonate of soda according to exactly the same method as had been 
employed in those lithia experiments which served for the calculation of the atomic weight. 

We had already ascertained that carbonate of soda can be fused in an atmosphere of 
carbonic acid without appreciable loss of carbonic acid (see next section). 

The results of these special experiments on carbonate of soda are given in the 
following table: — 





I. 


II. 


III. 


IV. 


C0 2 -pressure, 


60-0 


59-7 


29-2 


29-2 inch. 


Substance analysed, 


8-7906 


5-5292 


7-2228 


4-9794 grru 


C0 2 obtained, 


3-6446 


2-2922 


2-9960 


2-0669 


Percentage of C0 2 , 


41-460 


41-456 


41-480 


41-509 


Calculated value of Na 2 0, 


62-126 


62136 


62-076 


62-001 



Hence, mean value of Na^O = 62*0847, and Na= 23*042, which is a very fair approxi- 
mation to Stas's value 23*053. 

With these facts before me, I feel inclined to look upon Li= 6*89 as being at present 
the most probable value of this constant. 

Carbonate of Soda. 

The earlier experiments on this carbonate were made with a supply of Natrum 
carbonicum purissimum from Trommsdorff of Erfurt, which, as a qualitative analysis 
showed, really ivas very pure. But it subsequently appeared to me that it would be 
more satisfactory to work with an absolutely pure salt ; I accordingly, for the 
subsequent experiments, prepared a still purer salt in the following manner: — 

A quantity of purest soda-crystals ("purissimum" from Erfurt) was dissolved in 
water in a large nickel vessel, and almost, but not quite, neutralised with pure oxalic 
acid in the heat, to produce neutral oxalate, which separated out on cooling as a crystal- 
line powder. As neutral oxalate of potash is easily soluble in water, the trace of 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 455 

potassium which the original salt in all probability contained, could be expected to 
remain in the mother-liquor, with the rest of the impurities. The oxalate of soda was 
collected on a filter, washed cautiously with water, dried, and ignited in platinum. The 
residual carbonate was dissolved (in a platinum basin) in water, a trace of insoluble 
matter filtered off, the filtrate evaporated to dryness in platinum, and the residue 
dehydrated at a dull red heat without fusion, A salt prepared in this manner was used 
for the atomic-weight determinations reported on at the end of last section. 

Experiment I. — 8*462 grm. of (Erfurt) salt was heated in a current of pure and 
dry hydrogen for three hours by means of a powerful "Bunsen," which raised the 
temperature of the salt to considerably above its fusing-point; and then for other two 
hours over a gas blowpipe. The experiment was then stopped with the view of deter- 
mining the weight of the residue; but as the salt was found to have crept very perceptibly, 
besides having partially suffered volatilisation, the loss of weight observed is not worth 
stating. After this interruption the heating was resumed (the Bunsen and the blowpipe 
being used alternately), and continued until the crucible had been over the fire for in 
all ten hours, including five hours of blowpipe work. Of the product 2'0171 grm. were 
dissolved in water and the solution was made up to 18379 grm. This solution was 
preserved in a bottle under a vaselined stopper and parts of it used for the following 
analyses : — 

(1) As a preliminary test for caustic alkali, 50'434 grm. of the solution were mixed 
with excess of chloride of barium, the mixture allowed to settle in the absence of air, and 
the clear part examined. It was strongly alkaline ; an aliquot weight was titrated with 
fifth-normal hydrochloric acid, which showed that the substance contained 13 '5 9 per cent, 
of Na 2 0, uncombined with carbonic acid. 

(2) 50*399 grm. of solution, when decomposed with hydrochloric acid, &c, gave 
0-1871 grm. of C0 2 and 07857 of sulphate of soda. 

(3) 50-396 grm. of solution gave 0'1892 of C0 2 and 0-7850 of Na 2 S0 4 . Hence we 
have — 

(2) (3) Mean. Calculated* 

Carbonic acid, . . 33-83 34-20 34'02 34-07 

Soda, Na 2 0, . . 62-06 62'00 62-03 61-92 

Water, by difference, 4-11 3-80 3-95 4-01 



100-00 100-00 100-00 100-00 

These analyses consumed a considerable part of the product. The rest was put back 
into the crucible, and again heated in hydrogen over the blowpipe for five hours, with a 
view to seeing whether there would be any further diminution in the proportion of car- 
bonate. Unfortunately, however, when the crucible came to be opened, it was almost 
empty, and only 9 6 "2 mgrm. of the product could be scraped together for an analysis, 

* On the assumption that the substance is a mixture of 82 -16 per cent, of carbonate of soda, and 17 -84 per cent, of 
hydrate NaOH, equal (the latter) to 13 - 83 per cent, of Na 2 0, which agrees fairly with the determination by chloride of 
barium. 



456 



PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 



which gave 377 mgrm. of C0 2 = 39"18 per cent, and 132*8 mgrm. of Na 2 SO 4 :=60-31 per 
cent, of Na 2 0, leaving 0*52 for water, or : — 



Na 2 as carbonate, 

C0 2 , 

H. 2 as NaOH, 

Na 2 as NaOH, . 

Other Na 2 0, by difference, 



55-3} 

39-2 J 

0-51 

1-8 J 

3-2 



94-5 

2-3 
32 



100-0 100-0 



a very singular result, to which, however, as it was obtained by only one analysis of a very 
small quantity of substance, we must not attach too much importance. 

Thinking, rightly or wrongly, that the great loss of matter observed in this last 
experiment was owing more to the creeping-out than to volatilisation of caustic alkali, I 
endeavoured in the following experiments, Nos. 3 to 6, to prevent this by mixing the 
carbonate of soda operated upon with spongy platinum as explained above. This 
involved a slight modification in the method of analysis, consisting in this that the weight 
of substance analysed had to be determined indirectly, namely, by weighing the platinum 
cornet as it came out of the crucible (within a glass-stoppered wide test-tube), and again 
after expulsion of the carbonic acid and the complete removal of the soda-salt by 
exhaustive lixiviation with water. The spongy platinum was found to stick together and 
to the cornet, so that these operations offered no difficulty, although the washing process 
proved rather tedious. Experiment No. 3 was a mere rehearsal, to show the practicability 
of the method ; the results of Nos. 4, 5 and 6 are given in the following table : — 



(4) 



(5) 



(6) 



Time of heating over — 




(a) A Bunsen, 


18 hours 


(b) The blowpipe, . 


1 hour 


Dry Na 2 C0 3 used, . 


34784 


Product in cornet at the end, 




as analysed, . 


2-430 


Weight of C0 2 found, 


0-8826 


Na 2 S0 4 obtained, 


3-3908 


s by Calculation in Percentages — 


Carbonic acid, 


36-321 


Soda, Na 2 0, .... 


60-960 


Water, by difference, 


2-719 



Bunsen not used 


in 5 and 6. 


2 hours 


4 hours 


1-0342 


1-0068 grm 


0-8264 


04482 


0-2991 


0-1480 


1-1415 


0-6376 


36193 


33-021 


60-348 


62152 


3-459 


4-827 



100-000 100-000 100-000 

Assuming that the carbonic acid found is present as Na 2 C0 3 , and that the remaining 
soda and the water are united, as far as possible, into NaOH, the numbers, as they stand, 
would show that product No. (4) contains a small quantity of Na 2 0, uncombined with 
water, while (5) and (6) contain slight excesses of water. The anhydrous soda in No. (4), 
however, amounts to only about 0'3 per cent., and a very slight correction of the numbers 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 457 

suffices for causing it to vanish, as shown by the following entries for (4). Those for 
(5) and (6) are calculated from the uncorrected numbers. 

Percentage Compositions of — 





(4) 


(5) 


(6) 


Carbonic acid, . 


. 36-320 


36193 


33-021 


Soda, Na 2 0, as carbonate, . 


. 51-270 


51-090 


46-612 


Soda, Na 2 0, as NaOH, 


9-479 


9-253 


15-533 


Water as NaOH, 


2-747 


2-682 


4-502 


Other water, 


. Nil 


0-782 


0-332 



99-816 100-000 100-000 

Calculated percentage of Na 2 in 

No. 4, . . . . 60-749 ; instead of 60-960 as found. 

Of the four products analysed, only No. (5) is proved to have contained a little "free" 
water ; and its presence is only too easily explained, as having been absorbed on the way 
from the crucible to the balance. A mixture of carbonate of and caustic soda is of course 
highly hygroscopic. For the very same reason it is impossible to prove that the products 
did not contain even a trace of anhydrous oxide of sodium, beside NaOH. But the 
balance of probabilities is against this assumption. Assuming the absence of Na 2 to be 
proved, it would not follow that the caustic alkali is formed from the carbonate only 
by the chemical action of the hydrogen on the C0 2 ; thus : Na 2 C0 3 + H 2 = 2NaOH + C0 5 
and none of it by a mere decomposition of the Na 2 C0 3 into C0 2 and Na 2 0. The question, 
I thought, might be decided by a repetition of the experiments in an atmosphere of nitrogen 
instead of one of hydrogen gas, and I thus came to carry out the following four experiments, 
Nos. (7) to (10). In Experiments (7), (8) and (9) the modus operandi, apart from the 
substitution of nitrogen for hydrogen, was the same exactly as in Experiments (4) to (6). 
As the determinations of soda as sulphate ran away with a great deal of time, the last 
experiment (No. 9) was started before the analysis of product No. 7 could be calculated ; 
or else I should have found out sooner than I did what so clearly appears from the 
following reports, namely, that the hydrogen of the flame in this mode of operating 
diffuses through the platinum of the crucible, and produces a very tangible amount of 
NaOH. After some meditation on means and ways for keeping this hydrogen away from 
the alkali, I at last adopted (for Experiment (10) and later similar experiments) the method 
which I took occasion to describe on p. 443, &c, of this memoir, in connection with 
experiments on carbonate of lithia, i.e., I placed the carbonate of soda in a platinum 
boat, and heated it in a muffle within a porcelaiu tube through which a current of perfectly 
dry nitrogen was passing from beginning to end. In the first experiment of this series 
(No. 7) the nitrogen was made ex tempore, by passing a current of dry air through a 
column of red-hot copper wire gauze and thence direct into the crucible ; but it turned 
out that what was supposed to be nitrogen contained about 10 per cent, of oxygen. To 
my surprise the crucible at the end of the experiment was found to have been only very 



458 



PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 



slightly attacked by the alkali. In the subsequent experiments, the nitrogen was 
prepared by shaking air with a solution of pyrogallate of soda in one of the bottles of a 
Pisani gas-holder, the indiarubber tube connecting it with the other being clipped 
meanwhile. Our Pisani enabled us to operate on about 10 litres of air at a time; and 
24 grm. of caustic potash sticks, 10 grm. of pyrogallol, and 400 c.c. of water were 
found to be convenient quantities for that volume of air. In each experiment we had 
two Pisanis full of nitrogen ready, and both so connected with the apparatus that, as 
soon as one was emptied, the other could be substituted for it without loss of time. The 
nitrogen of the gas-holder was, however, not used as it came out, but first passed through 
a tower of sulphuric acid and pumice, and thence through a tube full of red-hot copper 
gauze, before it entered the porcelain tube. The results of Experiments 7, 8 and 9 are 
given in the following table : — 

Time of heating* 
Product obtained, 
C0 2 from it, . 
Na 2 S0 4 from it, . 



From these numbers we have in percentages : 



Carbonic acid, . 

Soda, Na 2 0, 

Water, by difference, . 



100-000 



(7) 


(8) 


(9) 


1-5 


2-5 


6 - 5 hours 


1-9978 


1-731 


2-0804 grm. 


0-8009 


0-689 


0-8385 


2-7096 


2-358 


2-8220 1 


mtages :■ — 






(7) 


(8) 


(9) 


40-089 


39-803 


40-304 


59-252 


59-509 


59-258 


0-659 


0-688 


0-438 



100-000 



100-000 



Or, by calculating the C0 2 into Na 2 C0 3 , and the water into NaOH, 





(7) 


(8) 


(9) 


Carbonate of soda, 


. 96-681 


95-991 


97-199 


Hydrate of soda, NaOH, 


2-933 


3-062 


1-949 


Oxide of sodium, Na 2 0, 


0-388 


0-948 


0-852 



100-002 



100-001 



100-000 



In Experiment (10) the weight of dry carbonate started with was 2*3407 grm.; the 
heating, after attainment of the highest temperature, was continued for two hours ; the 
loss of weight which the substance suffered was 0*1004 grm. Two analyses of the product 

were made ; the results were as follows : — 

I. II. 

Substance taken, 1*2564 0*941 

Carbonic acid obtained, 0-5107 0-3836 

Sulphate of soda obtained, .... 1-7027 1*276 

* Only the blowpipe used in all cases. 

t Mean of two closely-agreeing determinations. In this Experiment (9) for the first time carbonate of soda from 
oxalate was used. 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 



459 



Hence by calculation — 
Carbonic acid, 
Oxide of sodium, Na 2 0, 
Water, by difference, . 


I. 

. 40-648 

. 59-204 

0-148 


II. 

40-765 

59-238 

-0-003 


Mean. 

40-706 

59-221 

0-073 



100-000 



100-000 



100-000 



The 0'073 per cent, of water maybe put down as an observational error ; if we do so, 
and calculate all the Na 2 not carbonate as Na 2 0, we have : — 



Carbonate of soda, 
Oxide of sodium, 



98168 
1-759 

99-927 



This experiment clearly shows, that at the temperature at which it was made (which, 
according to a direct trial with beads of the metals, lay somewhere between the fusing- 
points of gold and silver), carbonate of soda does suffer dissociation to a very ap- 
preciable extent. And, as the temperature which prevailed in those experiments which 
were made in a directly heated platinum crucible, must have been higher, we must 
presume that in these also, the reaction Na 2 C0 3 + H 2 == 2NaOH + CO was accompanied 
by a simple decomposition of carbonate into oxide and carbonic acid. 

From Scheerer's results (see p. 440) it would appear that the dissociation-tension of 
carbonate of soda has an appreciable value even at temperatures not far above the fusing- 
point of the salt ; but it is not likely at such temperatures to come up to one atmosphere. 
So I thought, and to test my presumption, made the following experiment : — 

Experiment (11). — About 4 grm. of carbonate of soda (prepared from pure oxalate) 
was fused in a platinum crucible by means of a powerful Bunsen, and kept over this 
lamp for three hours, while a current of dry carbonic passed through the crucible. 
The carbonic acid was stopped shortly after the lamp had been withdrawn, and the 
contents allowed to freeze. The percentage of carbonic acid was determined twice with 
the following results : — 

Mean. 





I. 


II. 


Substance taken, 


. 1-2244 grm. 


1-9493 grm. 


Carbonic acid obtained, 


. 0-5112 


0-8093 


Percentage of C0 2 , 


. 41-75 


41-52 



41-63 

Pure carbonate of soda by calculation contains 41 '466 per cent. To another 
preparation subsequently made, the chloride of barium method for the determination of 
caustic alkali was applied by Mr Henderson, who obtained a negative result, although, 
according to his own synthetical results, less than 0*05 per cent, could easily have been 
detected. (Compare results given on p. 454.) Our preparation clearly was pure normal 
Na 2 C0 3 . It differed markedly from the ordinary preparation ; while the latter is very 
decidedly hygroscopic and perfectly opaque, our salt was almost transparent like glass, 
and devoid of all hygroscopicity. It struck me that carbonate of soda fused at the 



400 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

lowest sufficient temperature in carbonic acid would make a capital standard substance 
for the purposes of alkalimetry ; I therefore caused Mr Anderson to prepare a fresh 
quantity, and determine its degree of hygroscopicity by quantitative experiments. He 
fused a quantity of perfectly pure carbonate made from oxalate over a Bunsen for one 
hour in carbonic acid, and allowed the fuse to freeze in the crucible. It was then taken 
out, broken into small fragments, and two separate quantities of these weighed out, 
namely, one amounting to several grammes on a watch glass, and another in a stoppered 
preparation tube. The watch glass was allowed to stand uncovered in the balance 
case, and weighed from time to time. After forty-five hours it had gained 0*2 mgrm. 
After having been allowed to stand for other seven days, it had gained 4 '6 mgrm. 
The contents of the preparation tube during all the time had gained only 1 mgrm. 
The substance, as we see, is so little hygroscopic that, supposing it to be preserved in not 
too small fragments in a well-stoppered bottle, its weight could be relied on as remaining 
constant for a long time. 

At the very high temperature producible in a platinum crucible by means of a 
blowpipe, carbonate of soda loses carbonic acid even in an atmosphere of dry C0 2 , perhaps 
through rapid diffusion-in of H 2 . So at least it appears from the following experiment, 
which was made long before No. (11) : — 

Experiment (12). — 3 "1277 grm. of dry carbonate of soda (Erfurt purissimum) were 
placed in the hydrogen-crucible and heated in dry carbonic acid, first for two hours over 
a large Bunsen, then for other two hours over the gas blowpipe. The product was analysed 
with the following results — : 

27866 grm. gave 1*1497 of C0 2 and 3731 of Na 2 S0 4 , corresponding to 41 '258 percent, 
of carbonic acid and to 58*491 per cent, of oxide of sodium ; total = 99749. The deficit 
of 0*251 per cent., if not due to errors, must be put down as water. By combining the 
carbonic acid with what it needs of Na 2 to become Na 2 C0 3 , and calculating the rest of 
the Na 2 as NaOH, we have — 

Carbonate of soda, . . . . . 99-498 

Hydrate of soda, NaOH, 0-324 

Water, 0178 



100-000 



Normal carbonate of soda contains, by calculation, 41*47 per cent, of C0 2 and 58*53 
per cent, of Na 2 0. Unfortunately the chloride of barium method was not yet worked out at 
the time, so that the existence of caustic alkali in the substance was not proved directly. 

Experiments with Carbonate of Potash. 

A supply of " Kali carbonicum purissimum " from Trommsdorf of Erfurt, when 
tested qualitatively, turned out to be so nearly pure that it might safely have been used 
for the experiments as it stood ; but on the principle of good is good and better is better, 



HYDRATES AND CARBONATES OF THE ALK ALT-METALS, ETC. 461 

it was purified in the following manner : — A solution of the salt in water was poured into 
a solution of more than the calculated weight of tartaric acid in water acidulated with 
hydrochloric acid, the bitartrate allowed to settle, collected on a funnel over a small 
filter, and washed with small instalments of water, until the last washings were free of 
chlorine, when it was assumed that the small quantity of soda which the salt undoubtedly 
contained had passed into the filtrate ; the function assigned to the hydrochloric acid was 
to convert any calcium that might be present into chloride. 

The washed bitartrate was dried and heated in a covered platinum basin over a large 
flame, until all the organic matter was apparently destroyed ; the black mass was then 
treated with water, the charcoal filtered off, the filtrate treated with washed carbonic acid, 
until every trace of caustic alkali that might have been formed was sure to have been 
converted into, at least normal, carbonate, and then evaporated in platinum to dryness. 
The residue was made into a coarse powder, kept in a covered platinum vessel over a 
Bunsen at a temperature just short of the fusing-point of the salt, with occasional turn- 
ing over with a platinum spatula, and then preserved for the experiments. In the course 
of these it was observed that the caustic potash produced volatilised far more readily than 
the caustic soda had done in the previous trials with its carbonate, so that the outlet tube 
of the hydrogen crucible not unfrequently got clogged up, and had to be cleared by 
applying a special lamp and introducing a stout red-hot wire, to assist the hydrogen in 
blowing out the liquid alkali. 

Experiment I. — 1*67 14 grm. of dry carbonate of potash were mixed with spongy 
platinum, and the mixture, after having been put into a platinum foil cornet, heated 
within the gas-crucible in hydrogen for an hour and a half. Weight of product (as far 
as in the cornet after the heating process) = 0*4892 grm. Carbonic acid in it, 0*1422 ; 
sulphate obtained = 0*6474. Hence, by calculation, 

Carbonate of potash, K 2 C0 3 , . . . . . 91*346 

Oxide of potassium, K 2 0, ..... 9*286 



100*632 



The apparent absence of alkali hydrate, KHO, in the product causes me to look 
upon this analysis with suspicion ; I give the experiment for what it may be 
worth. 

Experiment II. — In this experiment 9*45 grm. of carbonate of potash were used and 
placed direct in the crucible, without spongy platinum. The blowpipe was again used 
from the first, and the heating continued for an hour and a half. The product was 
analysed twice : — 

Substance taken, ..... 
Carbonic acid obtained, .... 
Sulphate of potash obtained, 

VOL. XXXV. PART II. (NO. 12). 



(1) 


(2) 


1*835 


1-676 grm. 


0-5323 


0*4861 


2-3468 


2*1452 




4 G 



462 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 
Hence by calculation — 

(1) (2) Mean. 

Carbonic acid, . 29'008 29 '004 29-006 

Oxide of potassium, 69163 69-219 69-191 

Water, by difference, . ... ... 1-803 



100-000 
If we calculate the C0 2 as K 2 C0 3 , and the rest of the alkali as hydrate, there is a small 
excess of water left, thus — 

Carbonate of potasb, ..... 91153 

Caustic potash, KHO, 8'389 

Water, 0-458 



100-000 
Experiment III. — An unrecorded weight of carbonate of potash was heated in the 

platinum crucible in nitrogen for an hour and a quarter, by means of the blowpipe. 

No spongy platinum used. 

1-4516 grm. of product gave 0-4508 of C0 2 , and 1-8337 of K 2 S0 4 . Hence, by 

calculation — 

Carbonate of potash ...... 97'595 

Hydrate, KHO, 2-115 

Water, 0-290 



100-000 

Experiment IV. — An unrecorded weight of carbonate of potash was placed in a boat, 
and heated in a muffle, within a porcelain tube, in a current of nitrogen, for two 
hours. 

0-834 grm. of product gave 0'2645 of C0 2 and 1*055 of K 2 S0 4 ; or 

Carbonate of potash, ...... 99 - 663 

Oxide of potassium, . . . . . . 0457 



100120 

In another sample, Mr Henderson found, by means of the chloride of barium method, 
0*49 per cent, of K 2 0. 

Experiment V. — A quantity of carbonate of potash was heated in the platinum 
crucible in a current of dry carbonic acid, first for one hour by means of a small Bunsen 
and then for an hour and a half with a more powerful lamp. The residue, however, was 
not analysed, because it exhibited a brownish colour, as if the organic matter, in the 
ignition of the bitartrate, had not been quite completely destroyed. A similar colour had 
been noticed in the residue obtained in Experiment III., but not in that of Experiment IV., 
although the same preparation had been used throughout. I did not think, and do not 
now, that the remnant of organic matter (if there was any) amounted to more than a 
minute trace ; yet to make sure that in this experiment at least there should be no 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 463 

impurity of any kind or magnitude, a quantity of the carbonate was fused in instalments 
over the blowpipe in a platinum crucible, the fused salt dissolved in water, the filtered 
solution treated with carbonic acid, to saturate any caustic alkali and then again evapor- 
ated to dryness in platinum, and dehydrated without fusion. The experiment was then 
repeated with the purified salt, i.e., a quantity of it was heated in carbonic acid, first for 
one hour with a small, and then for an hour and a half with a larger Bunsen, and 
allowed to cool in C0 2 . The product this time had no brownish tinge about it ; some 
parts of it had that glassy appearance which I had noticed in the corresponding kind of 
carbonate of soda. It was analysed twice, with the following results :■ — 

(1) (2) Mean. 

Substance, . . . 1-4442 1-9376 

Carbonic acid obtained, . 04594 - 6169 

Percentage of carbonic acid, 31-809 31-838 31-824 

Pure K 2 C0 3 contains, by calculation, 3L821 per cent.* 

Mr Henderson applied the chloride of barium method, and found no caustic alkali. 
The experiment is interesting chiefly as showing how perfectly normal fused K 2 C0 3 can 
be prepared. 

Experiments with Carbonate of Rubidium. 

Let me state at once, that the word " Rubidium " in this heading means " a mixture 
of rubidium and caesium, consisting chiefly of the former." The research on the " Chloro- 
platinate Method for the Determination of Potassium," &c, which I carried out sometime 
ago with Mr MAcARTHUR,t had left me in possession of a number of residues, from which 
such " rubidium " could be and was extracted by known analytical methods. The 
extracted rubidium was converted into normal sulphate, with the view, originally, of con- 
verting this into hydrate by means of baryta water, and the hydrate into carbonate by 
combination with carbonic acid. But caustic rubidia must be presumed to attack glass 
very strongly, and we had not at the time a large enough platinum vessel to spare, which 
would have enabled us to avoid contact of the caustic rubidia with glass. I therefore 
caused Mr Henderson to try and convert the sulphate into carbonate more directly by 
treatment with carbonate of baryta, water and carbonic acid, which last-named reagent 
was intended to convert the alkaline carbonate into bicarbonate as quickly as it is pro- 
duced, and to act as a solvent for the carbonate of baryta. To my surprise this very 
plausible method would not work ; the process stopped long before the whole of the 
sulphuric acid had assumed the form of baryta salt. 

I therefore proposed to substitute oxalate for the carbonate of baryta, and was glad 
to find that the decomposition this time went on as desired. To produce an oxalate of 
baryta sure to be free of alkali, this salt was made by addition of a solution of pure oxalic 
acid to its calculated equivalent of pure baryta water, and then rendering the mixture 

* If K = 39-136, as found by Stas. If K = 39-00, the percentage becomes 31-884. I am far from taking it for 
granted that the close agreement of our number with Stas's is not a mere accident. 
+ Transactions of the Royal Society of Edinburgh for 1887, p. 618. 



464 PROFESSOR W. DTTTMAR ON THE BEHAVIOUR OF THE 

slightly alkaline with more baryta. The precipitate was allowed to settle, washed, and 
dried. After a preliminary trial 7 '8 6 grm. of the sulphate of rubidium were dissolved 
in about 400 c.c. of hot water. 6'6 grm. of oxalate of baryta were now added, and the 
mixture was allowed to stand for three hours with frequent agitation. As soluble 
sulphate was still present, the precipitate was filtered off, and some more of the oxalate 
added, which removed the whole of the S0 3 in other two hours. The filtered solution was 
evaporated to dryness and ignited gently in platinum. The residue was dissolved in 
water, a small precipitate (BaC0 3 ) filtered off, and the clear solution again evaporated to 
dryness in platinum and ignited. The total quantity of carbonate obtained in this and 
the preceding pioneering operations amounted to about 10 grm. Of this small supply 
more than one-fifth was lost in the first experiment. 

Experiment I. — 2*04 grm. of carbonate of rubidium were placed in the gas-crucible, 
and heated in a current of hydrogen over the blowpipe for one hour. When the 
crucible was opened, it was found to be empty : all the salt had volatilised ! 

Experiment II. — Seeing that the rubidium salt is so very volatile I took care this 
time to substitute an ordinary Bunsen for the blowpipe. About 2 grm. of the carbonate 
were operated upon, and the heating in hydrogen continued for one hour. The aspect of 
the residue showed that the salt had been in a state of fusion. For the analysis of the 
product, 1*4336 grm. of it were decomposed with hydrochloric acid in the usual apparatus, 
and the C0 2 collected and weighed. It amounted to 0*2561 grm. or 17 "864 per cent, of 
the substance analysed. The chloride solution produced was evaporated, finally in a 
small platinum basin and the residue dried at 150° for two hours; the salt was then 
moistened with water and again dried at 150°, to expel the uncombined hydrochloric 
acid, and the drying process continued until the weight of the residue was constant, at 
1-5238 grm. 

This residue (supposed to consist of only chloride of rubidium), when dissolved in 
water, left a residue which was filtered off, ignited, and weighed. It weighed 3 mgrm. 
The filtered solution was diluted to 201*465 grm., and divided into two portions. In one, 
100*589 grm., the chlorine was determined with nitrate of silver ; weight of chloride of 
silver obtained 0*9036 grm., corresponding to 0*22343 of chlorine or 0*4475 for the 
whole. The other portion, 201*465 — 100*589 grm., was treated with sulphuretted hydro- 
gen in the heat, which caused the formation of a small black precipitate. This precipitate 
was filtered off, ignited, and weighed, and found to amount to 3*5 mgrm., or 7*0 for the 
whole. The filtrate was evaporated to dryness, and the residue dried at 150° until 
constant in weight, at 0*760 grm., or 1*5178 for the whole. But this probably included 
a little sulphur from the excess of sulphuretted hydrogen, which had not been removed. 
The platinum-free chloride gave 0*8964 grm. of chloride of silver, corresponding to 0*22165 
of chlorine or 0*44267 for the whole. The 3 mgrm. of insoluble matter from the crude 
chloride were unfortunately not examined qualitatively ; the ignited sulphuretted hydro- 
gen precipitate consisted of metallic platinum. This platinum, no doubt, was present in 
the fuse in the form of Pt0 2 + xK 2 0, produced, in spite of the hydrogen, at the expense of 



HYDKATES AND CARBONATES OF THE ALKALI METALS, ETC. 465 

the carbonic acid of the carbonate of rubidia, thus : — 2E 2 OC0 2 + Pt = Pt0 2 + 2R 2 4- 2CO. 
If this is so, then the insoluble matter obtained from the crude chloride must be presumed 
to have included chloroplatinate of rubidium. Supposing it contained nothing else, 
the 3 mgrm. of ignited precipitate consisted of the mixture PtCl 2 E 2 , and indicated 2*8 
mgrm. of Pt0 3 E 2 in the substance analysed. 

The platinum-free chloride contained less chlorine than the crude by 4 '8 mgrm., which 
agrees well with the 5*08 demanded by the 7 mgrm. of dissolved platinum found for con- 
version into PtCl 4 . The 3 mgrm. of insolubles correspond to 4*0 mgrm. of PtCl 6 R 2 . 
Deducting these, and the 11 '8 mgrm. of dissolved chloride of platinum, we have 1508'0 
mgrm. for the weight of the chloride of rubidium present as such or as chloroplatinate in 
the soluble part of the crude chloride. But the chlorine in this, as we saw, amounted to 
0-44267 = 12-485x01 mgrm. Hence we have RC1= 1508*0 : 12-485 = 120-78, and 
R= 85*33, which agrees with the equivalent of unmixed rubidium. At this, however, 
we need not wonder, as the caesium carbonate may well be presumed to have volatilised 
during the ignition process. The carbonic acid found amounts to 256*1 = 11-641 x 22 
mgrm. The total platinum found corresponds to 9'705 mgrm. of oxide = "08 5 x 0'5 Pt0 2 
mgrm. 

Summing up, we have : — 

In Milligrammes, 
absolute weight. 
1166*54 
256*10 
9*705 
6*957 



Total rubidia, E 2 0, . 


In Milligramme 
equivalents. 
12-499 


Carbonic acid, 


11-641 


Oxide of platinum, . 


0-085 


Water, as EHO, 


0-773 



Total, . . 1439*30 

The substance analysed weighed, .... 1433*60 

Excess found, . . . . . . .5*70 

which just about corresponds to the " water " found by calculation. It does not follow 
that the water was absent, i.e., the surplus rubidia all present as platinite or anhydride. 

To avoid unnecessary repetitions, let me state that in the following experiments (III., 
IV., and V.), when the carbonate was heated in dry nitrogen or carbonic acid, the product 
was platiniferous, as in the case of Experiment II.; and had to be analysed by a similar 
method. 

Experiment III. — 1*308 grm. of carbonate of rubidia were heated in the gas-crucible 
over a Bunsen lamp in dry nitrogen for one hour. 1*0383 grm. of the product gave 
1093*7 mgrm. of crude chloride, yielding 5'5 mgrm. of (ignited) insolubles. The chlorine 
in the soluble part amounted to 318*02 mgrm; the chlorine in the platinum-free chloride to 
316*99 mgrm.; difference, 1*03 mgrm. But the platinum found in the solution was 5*2 
mgrm., demanding 3'78 mgrm. of chlorine for its conversion into tetrachloride. Hence 
one at least of the two chlorine determinations cannot be quite correct, and, as we do not 
know which it is that is at fault, we will deduct these 3*78 mgrm. of chlorine from the 
318-02 found in the crude chloride, and adopt the mean between the remainder and the 



466 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

316*99 found directly for the platiuum-free chloride as representing the true value for 
the latter. This mean is 315 -61 = 8 '9018 x CI mgrm.; so many equivalents of rubidia 
were in the soluble part of the chloride. In addition thereto, if we view the 5*5 mgrm. 
of ignited insolubles as consisting of PtCl 2 R 2 , we have, for this residue, 2*33 mgrm. of R 2 
and 2*85 mgrm. of oxide of platinum. The carbonic acid amounted to 197*0 mgrm. 
Hence we have — 

In Milligramme In Milligrammes, 

equivalents. 
Eubidia, B, 2 0, . . . 8-9267 

Carbonic acid, . . . 8*9547 

Oxide of Platinum, . . 0-0780 



absolute weight. 
8331 * 
197-0 
8-9 


. 1039-0 
. 1038*3 



Total, 

Substance analysed, 

which agrees satisfactorily, but the equivalents of acid and of base do not balance each 

other. Remembering that the number for the carbonic acid by itself is uncertain by 

± half a mgrm., and consequently the corresponding quotient by ±0*023, the analysis 

may be read as showing that the product was normal carbonate of rubidia mixed with 

oxide of platinum. 207 mgrm. of the product were handed to Mr Henderson for the 

determination of the free base by the chloride of barium process. He found 0*07 mgrm. 

of Rb 2 0, or 0*03 per cent. But this corresponds to only a drop of the very dilute acid 

used ; hence what he really made out was that the product contains a just perceptible 

trace of free base. 

Experiment IV. — Not feeling quite sure of the result of the last experiment, I 
repeated it with 1*838 grm. of carbonate, with this difference, however, that I handed the 
whole of the product to Mr Henderson for the determination of the free alkali by the 
chloride of barium method. He found, in 1*785 grm., 2*79 mgrm. of Rb 2 0, or 0*156 per 
cent. This, of course, includes the platinite, which must be presumed to have been present. 

Experiment V. — In this experiment 2*356 grm. of the carbonate were heated in the 
gas-crucible over a Bunsen in a current of dry carbonic acid for one hour. 1*3382 grm. 
of the product, when treated with hydrochloric acid, gave 252*0 mgrm. of carbonic acid 
= 11*455x22 mgrm., and 1409*6 of crude chloride, containing 2*0 mgrm. of (ignited) 
insolubles, and, in the soluble part, 411*52 mgrm. of chlorine = c', and 1*0 mgrm. of 
platinum, demanding 0*73 mgrm. of chlorine for its conversion into tetrachloride. 
The platinum and HCl-free solutum contained c" = 410*70 mgrm. of chlorine, 
= 11*548 x CI. Now, c' — c" = 0'82 mgrm., which is a sufficient approximation to 0*73. 
The 2*0 mgrm. of insolubles (if PtCl 2 R 2 ) correspond to 2*65 mgrm. of chloroplatinate ; 
hence the weight of the platinum-free chloride, calculated from the crude, is 
1409*6- (2*65 4- 1*73)= 1405*2 mgrm.; hence RC1= 1405*2 : 11*584 = 121*31 ; whence 
R = 85*85 and J>R 2 = 93*85. The 2 mgrm. of ignited insolubles correspond to 0*83 
mgrm. of rubidia = 0*009 eqq., and to 1*04 mgrm. of oxide of platinum. The total 
platinum oxide amounts to 2*20 mgm., = 0*019 x ^Pt() 2 . Hence we have — 

* R = 85 , 33, as found in Experiment II. 



i Milligramme 

Equivalents. 

11-593 


In Milligrammes, 
absolute weight. 
1088-0 


11-455 


252-0 


0-019 


2-2 



HYDRATES AND CARBONATES OF THE ALKALI-METALS, ETC. 467 



Rubidia, R 2 0, 
Carbonic acid, . 
Oxide of platinum, 

Total, . 1342-2 
Which is 4 mgrm. more than the substance analysed, . 1338*2 

From the quotients it would appear that the substance contained 0*138 eqq. of rubidia 
not carbonate in 11*593 eqq. of total rubidia; yet the chloride of barium test showed no 
free alkali at all. 

It is rather remarkable that fused carbonate of rubidia at a red heat attacks platinum, 
even in an atmosphere of hydrogen, nitrogen, or carbonic acid. Id the case of hydrogen 
an appreciable quantity of caustic rubidia is formed independently of the action of the 
platinum ; but whether a similar assertion holds in the case of nitrogen or carbonic acid, 
my analyses are not sufficiently exact to show. I hope, one day, to repeat these experi- 
ments in gold vessels, and to thus settle the question. 

SUMMARY. 

I. Regarding the Hydrates. 

The hydrates of barium and lithium, if kept at a red heat in an atmosphere of 
hydrogen, quickly lose their water, and become pure monoxides, obviously a plain case 
of dissociation in which the hydrogen plays no chemical part. The result, no doubt, 
would be the same in an atmosphere of nitrogen, or in a vacuum. The behaviour of the 
hydrates of the alkali-metals, properly so called, still remains to be ascertained. 

II. In regard to the Carbonates. 

1. Carbonate of baryta, if kept at a red heat in hydrogen, gradually loses its carbonic 
acid, and at last leaves a residue of pure oxide. In an atmosphere of nitrogen no 
decomposition takes place. 

2. Carbonate of litliia, at a red heat in hydrogen, loses its carbonic acid very slowly, 
but at last completely, and is reduced to Li 2 0. 

In an atmosphere of nitrogen, the same reduction takes place, but it proceeds still 
more slowly. A residue obtained after ten hours' heating had, approximately, the 
composition Li 2 + Li 2 C0 3 , indicating the existence of a relatively stable basic carbonate. 

In an atmosphere of carbonic acid, of any pressure from 30 to 60 inches of 
mercury, normal carbonate remains. The composition of this carbonate shows that the 
atomic weight of lithium is less than 7*00, the value demanded by my analyses being 
6*98±0*01. 

3. Carbonates of Soda and Potash. — These carbonates, if kept at a red heat in an 
atmosphere of carbonic acid, remain unchanged ; in an atmosphere of nitrogen or 



468 PROFESSOR W. DITTMAR ON THE BEHAVIOUR OF THE 

hydrogen they lose carbonic acid with formation of oxide, R 2 0, or hydrate, RHO. In the 
case of nitrogen, if contact of the substance with flame-gases be absolutely avoided, the 
change proceeds very slowly. A product obtained from the soda-salt had the composition 
Na 2 CO 3 + 0'0306 Na. 2 0. A product obtained from the potash-salt had the composition 
K 2 C0 3 + 0-00673 K 2 0. 

In an atmosphere of hydrogen the decomposition goes further, with formation of 
only hydrate, RHO. The most highly caustic products obtained had the following 
compositions : — 

In the case of soda, Na 2 C0 3 + 0'3332 Na 2 O.H 2 0, or very nearly 3 Na 2 C0 3 + lNa 2 O.H 2 0. 

In the case of potash, K 2 CO 3 + 0*1134K 2 O.H 2 O ; or approximately 9K 2 C0 3 + 
1K 2 0.H 2 0. 

These formulas, however, have a purely arithmetical meaning. We have no reason to 
assume the existence of any such hydric carbonate as a chemical compound. That we 
never obtained a residue consisting of unmixed hydrate is easily explained by the 
relatively great volatility of the latter. Imagine, at a certain stage of the change, the 
crucible contained x mgrm. of R 2 as carbonate and y mgrm. of R 2 as hydrate, then in 
the next second, a small weight of fresh hydrate proportional to x will be produced, and 
a small weight of the already formed hydrate proportional to y will be volatilised, so 
that, at the end of that second we have x' mgrm. of R 2 as carbonate and y' mgrm. of 
R 2 as hydrate, where, approximately at least, x' = x — Jcx and y' = y + kx — hy, where k 
and h are constants whose meaning is obvious. The ratio x : y or the weight of R 2 
present as carbonate per unit-weight of R 2 present as hydrate, passes from x : y to 

X ^ ™ d y> = y(l+Z*-h)- 



x' 



Supposing h is less than k, the denominator in the factor of — is greater than the 

numerator, x' : y' is less than x : y and this diminution will continue until at last the 
carbonate present per unit of hydrate will be nil and the residue consist of pure hydrate. 
This case is illustrated in our experiments on lithia, except that there the anhydride 
Li 2 made its appearance. 

Supposing noiv h is greater than k, then, in the earliest stages of the process, k- is 

very great, so that the factor is less than unity ; the ratio x : y becomes less and less, 

but as it does so the denominator in the factor becomes less and less likewise and at 

last equal to 1 — k, and the ratio x : y becomes constant. This will occur as soon as x : y 

has assumed that value at which 

-.x 

or 



when 



k + k- - 


-h = 


= o, 


y 






Ml + — 

\ y 


■8- 


= o, 


x h 






y~k~ 


-i 





or 



HYDRATES AND CARBONATES OF THE ALKALI- METALS, ETC. 469 

In the most highly caustic product, I obtained from carbonate of soda - was equal 

to 3, hence in that experiment we may have had 

h h 

3 = T — 1 , or y = 4 , or h = 4k . 
k k 

But both k and h are functions of temperature ; both must be presumed to increase when 

the temperature increases, but h may increase faster than k. If so, h : k and consequently 

x:y, i.e., the proportion of carbonate in the residue increases as the temperature rises. 

This would explain the apparent anomaly which presented itself in the experiment with 

carbonate of soda, detailed on page 028 ; provided the result was not owing to unobserved 

errors. 



VOL. XXXV. PART II. (NO. 12). 4 H 



( 471 ) 



XIII. — On the Determination of the Curve, on one of the Coordinate Planes, ivhich 
forms the Outer Limit of the Positions of the Point of Contact of an Ellipsoid 
of Revolution which always touches the Three Planes of Reference. By G. 
Plare, Docteur es-Sciences. 

(Read 1st July 1889.) 

In a former paper {Trans. Roy. Soc. Edin., vol. xxxiii. part ii. p. 465) the ellipsoid 
under consideration was supposed to have its three axes different from one another in 
length. I consider now the more simpler case of an Ellipsoid of Revolution. The 
same notations as in the former paper shall be made use of, and for the establishment of 
some of the results which form the basis of the present paper, reference must be made to 
the former. 

The condition (§ IV. of the former paper) 

i/,s = 
leads to the consequence 

This is to be satisfied for all points on the limiting curve. 
If we put generally for any point : 

K = Nyjs'j^'k = iM 1 + jM/ + &Mj", 

we get by the expressions (§ VI.) of \f/j , \p'k : 

Mj = R 1 R 2 -Q 2 
M/ =P 1 Q -P 2 R 2 
M/' = P 2 Q -PA- 

In the former paper we did not show the manner in which the coefficients PiP 2 , &c, 
are obtained, we may therefore be allowed to show the method for one or two of them; 
as, for example, 

P x = Si\Js 'j = Sjijsi . 
By the expression § III. we have 

= --Si 2 d>-H-J-Sijd>- 1 j - -SiU-ik 

u r- v jy j w -r 

+^(v^->i + < r ^)+£± i siv-'i 



= -i^-i + W^' 



w " ' u 



Now 

V^- 1 i = Vjk^- 1 i = jSktp- 1 i — kSj^- 1 i . 

VOL. XXXV. PART II. (NO. 13). 4 I 



47'J DR G. PLARR ON THE DETERMINATION OF 

With the definitions (§ II.) of a?!, y u z x : 



This gives then 







i • J . & 

\//-i = — - 03, + — #, 






+ -V,— z. 






1 v u 


(- 


-Shfn) 


or P„ = -i - — * 



and so on. 

In the case of Q, R 1? R 2) we make some further transformation. We have 

u 3 Q = v?x x — y^z Y 

[it is to be remembered that </> is self-conjugate, therefore we have replaced 

a5i = S/^ _1 A by Sk^j] . 



Applying the formulae 



V^- 1 X0-V = ^0VX A * 



(where w = the first coefficient in the cubic relative to <f>), we get 



w 3 Q = — Sj0/c. 
m "^ 



Likewise in Ri, R 2 we have respectively 



Thus 



(yf - w 2 u 2 ) = SkiYf-^-H = -Sj0j 



1 m *' w w 



% 3 R, = -Sfc0ife+tt 4 -^ 



In § XIII. of the former paper we have stated the formulas 



(i) y-a, 

(2) z=w-^l 

v u 



P namely being supposed in the plane of j, Jc, we have assumed there 

p =jy + kz 



THE CURVE ON ONE OE THE COORDINATE PLANES. 

The last two terms in R 1} E 2 , may now take the forms respectively 

—[w — JD >) = u 4 — 

w\ u / w 

Vt( v -h\ =u &. 

v\ U/ V 

So that we have also 

u3R 1 = §M+^? 
1 m w 

The expression of M 1 now becomes 



m 



_/ SJ0j tt%\/ S&0fc u 4 y\ S 2 j<ftfc 



\ m w /\ m v / 

= —2 [Sj0is/c^ - s?0&s%'] 



771" 



m 



U m w m J 



But the first term is 



vw 



= — 2 S.Vi/vV0%' 



w 



m- 



Si. m(j>~ 1 i= . 



m 



viv 



Substituting this, and multiplying the whole equation by — ^ , we get 



u 4 vwM 1 = 



vw 
m 

+ u 2 wy *■-'-' 



-+-U2;- 
m m 



] 



473* 



Of course, if M a = 0, the second member must vanish, the factor vtvw which was intro- 
duced being never susceptible of vanishing ; in other words, the centre of the ellipsoid 
being unable to coincide with one of the three planes of the octant. 

We may make further transformations before introducing the hypothesis 6 2 = c 2 . 
Squaring the expressions of y, z, under the form, 

u(y-v) = -z 1 
u(z-w)=—y 1 , 



v 2y2 _ 2u 2 vy = z 2 — u 2 v 2 



we get by the first 

the second member being 

Sitjt-ySjf-H - Si(p-HSj<j>- l j = SijV^-H^j 



= — Skd>Jc . 
m 



474 



DR G. PLARR ON THE DETERMINATION OF 



Hence 



The equation in Mx becomes 



' u 2 y 2 — 2u 2 vy = — Sk<bk 



u 2 z 2 — 2u 2 wz = — Si chj 



Or, 



0=-™ 
m 

+ u 2 [wy(u 2 z 2 — 2u 2 wz) + vz(u 2 y 2 — m 2 vy)] 
+u % yz. 

vw 



= 



m 



+ u 2 yzTu 2 wz — 2 -uPw 2 + u 2 vy — 2u 2 v 2 ~\ 

Assembling the terms in u 2 , v 2 , w 2 , and remarking 

u 2 + v 2 + w 2 = 2Si<j>-H = 2a 2 = l x 

(l t becoming later = a 2 + 2b 2 ), and having 

u* - 2u 2 w 2 - 2u 2 v 2 = u\u 2 - 2(v 2 + w 2 )] 
= u 2 [u 2 -2(l 1 -u 2 )] 
= u 2 [3u 2 -2^], 



we get 



771 



(3) +u 2 yz[u 2 (vy+wz)+u%3u 2 -2l 1 )] . 

The expressions (1) (2) give 

2u\vy + wz) = u 2 (y 2 + z 2 ) 



-ksj<pj+sk<pk). 



But the last term 



Putting y 2 + z 2 = r 



m 



we get 



+ -(S^+m 2 ), as ESi0i = s(+- 2 j=- 



2u 2 (vy + wz) = v?r 2 -\ — ($i<pi + m 2 ) 



m„ 



and the equation in Mj becomes 



m 



0=- 



2vw 



m 
r 



+u 2 yz 



u 2 r 2_j (Sid)i+m„) 

m 

+ w 2 (6u 2 -4£ 1 ) 



THE CURVE ON ONE OF THE COORDINATE PLANES. 475 

We have now two equations, the one giving the sum of vy and ivz, the other giving 
(when multiplied by yz) the product vy x wz. 

Let us put for abbreviation, 

u V 2 = t 

y 2 z 2 = s 
— (S^+m 2 ) = L 



we have 



Then we may put 



where 



vy+wz = — 2 (t + L ) 

vy x wz = — — (t + L + Lj) . 






We will now deduce two other expressions of v, w, from the equations (1), (2). But 
for this we must now have recourse to the hypothesis 

b 2 = c 2 , 

which constitutes the ellipsoid as a figure of revolution, the axis having the direction of a. 
In § I. of the former paper we have assumed by definition 



Making b 2 = c 2 , and remarking 
we get 

As 

becomes 



, aSaca f3S8w , ySyw 
/SS/3(o + ySyo> = — w — aSaw , 



^co = aSaa)^— 2 -pj- 



'p' 



1 

m= — 



a 2 b 2 c 2 

1 
~a 2 ¥ 



Likewise we have 
From these we deduce 



^ = aSa<ob 2 (a 2 - b 2 ) + coa 2 b 2 . 
<f>~ 1 co = aSaa)(a? — b 2 ) — wb 2 . 
^^[^co + ^ + b 2 )]. 



m 



476 DR G. PLARR ON THE DETERMINATION OF 

Hence (as &j<f>~\j = r, &c.) 

u 2 y 2 — 2u 2 vy = b 2 [w 2 — a 2 — b 2 ] 

u 2 z 2 — 2u 2 wz = b 2 [v 2 —a 2 — b 2 "] . 
But 

w 2 —a 2 —b 2 = l 1 —u 2 —v 2 -a 2 — b 2 ; 
as 

l l = "La 2 = a 2 + 2b 2 
therefore 

w 2 — a 2 — b 2 = b 2 — u 2 — v 2 

v 2 — a 2 —b 2 = b 2 — u 2 — w 2 . 

Putting every term into the first members of the preceding equations, and ordaining, we 

get the quadrics, 

b 2 v 2 - 2u 2 yv + [u 2 y 2 + b%u 2 - b 2 )] = 

b 2 w 2 - 2u 2 zw + [u 2 z 2 + b 2 (u 2 - b 2 )] = . 

Hence in resolving as to v, iv, 

b*v =u 2 y + Y | 

b 2 w = u 2 z+Z J ' 

where 

Y2 = u y _ &[ U Y + &2( u 2 _ &2 )] t 

Z 2 = u% 2 - b 2 [u 2 y 2 + b\v? - &)] , 
or also 

Y 2 = {u 2 -b 2 ){u 2 y 2 -¥), 

Z 2 ={v?-b 2 )(v?z 2 -¥) . 
Eliminating v, w from both of their values we get 

2y{u 2 y + Y) = ~[t + \ + 2n 2 ^\ 

20 (u 2 z + Z) = ^ [t + L - 2u 2 R] . 

4u 4 R 2 = R' 2 ,2u 2 R = R, 
R' 2 = (t + L ) 2 - 8mu 6 s(t + L + Lj ) 

4w2Yi/-6 2 R'=6 2 (i + L )-4uY . . . (I.) 

\u 2 Zz + 6 2 R' = b\t + L ) - 4u 4 2 . (II.) 

L o = -(S^+m 2 ), 
-Si0i = 6 2 [Si0 - H + i\a 2 + b 2 )] 



Putting 
namely, 
we get 

We have for L ( 
and as 



m 



= b 2 [u 2 -(a 2 + b 2 )] 
= ( - a 2 V)2 ( - ^\ = ¥ + 2a 2 b 2 ., 



THE CURVE ON ONE OF THE COORDINATE PLANES. 477 

we have 

L = 6 2 (u 2 +a 2 ). 

The equations (I.) (II.) contain the three elements y, z, and u 2 alone. 

To get rid of the three radicals R', Y, Z, we must have recourse to squaring twice 
over. 

[(If we had, by a more direct process, operated at once on (1), (2), (3), in order to 
get rid of the three radicals u, v, w, we would not have the advantage of knowing that 
v and w were of the forms respectively 

E+C, E-C. 
C being a radical)]. 

After squaring both members of (I.) we pass the radical YR' into the second member, 

the rational parts into the first, and we get 

16w 4 Yy-16uy ) = 8u 2 Yyb 2 K; 
+ ¥R' 2 -b\t + L Q ) 2 
+ 8 U yb%t + -L ) 
or 

16uY(Y 2 — uV) ) = 8u 2 yb 2 YK , 

+ 8uyb%t+L ) 
or by the values of Y 2 , R /2 : 

16%y ( - b 2 )[u 2 y 2 + b\u 2 - ¥) ) = 8u 2 yb 2 YK . 

-8mu 6 s(t+~L +~L 1 )b i 

+ 8uyb*(t+L ) 

We may suppress the factor 8u 2 yb 2 , by encroaching on s — y 2 z 2 . Thus putting 

Y' = - 2u\v?y 2 + b\u 2 - b 2 )] 

-muV6 2 (<+L +Li) 

+ u 2 (t+L ) 



we get 

where we put also 

As we have 

we replace, in Y r , 

Then as 



yY'=YK 
zTl = -ZR'. 

Z' = - 2u\u 2 z 2 + b\u 2 - 6 2 )] 
-mu^^+Lo+LJ 
+u 2 (t + L ). 

t = u 2 y 2 +u 2 z 2 

u 2 y 2 by t — u 2 z 2 . 

L = b 2 (a 2 +u 2 ), 

Y' = u 2 r— 2t + 2u 2 z 2 — 2b\u 2 - 6 2 )~| 
L+t+b 2 (a 2 +u 2 ) J 

-muV^+Lo+Li); 



478 DR G. PLARR ON THE DETERMINATION OF 

or grouping the terms in z 2 together 

Y' = u\ - 1 - b 2 u 2 + b 2 (a 2 + 2b 2 )] 

+ u*z 2 [2-mb 2 (t + L +L 1 )] . 



and as 



put this=L 2 . 
Thus we have, 



a 2 + 2b 2 = L : — mb 2 = -7?r„ 
a~b- 

+^[2a 2 b 2 + (t + U+^]. 

2a 2 b 2 + L = 2a 2 b 2 + b 2 a 2 + b 2 u 2 
L^w 2 ^ 2 -^) 
2a 2 b 2 +L + L 1 = Sa 2 b 2 +u 2 (b 2 -U 1 ) + 6u\ 

Y' = u 2 [b 2 (l x -u 2 )-q 
Z' = u 2 [6 2 (^-u 2 )-^] 



where 



L 2 = 3a 2 & 2 -u 2 (4a 2 +76 2 ) + 6u 4 . 
At the end of this paper we will show that we have identically 

M 1 _M 1 '_M/'_ 

— — xN j 

1 9 3 

where c^, a 2 , a 3 are the scalar coffiecients of a decomposed parallel to i, j, k, so that 

a = ia, +ja 2 + Jca 3 . 
The equations M/ = 0, M/' = give therefore no new relations besides those which we 
have deducted from (1), (2), (3). We may incidentally also remark here, that by the 
geometrical nature of the question we have identically 

because the point of contact (or of tangence) cannot change when the rotation-axis 
coincides with the axis of figure. 

We remark that Y', 71 are of the forms 

Y^Xo+3%! jX = u 2 [fc 2 (Z x -u 2 )-/] 

Z^Xo + ^j-^ 6 [x x =^(L 2 +t) s 

and remembering 

Y 2 = (u 2 y 2 -b i )(u 2 -b 2 ) 
Z 2 = (u 2 z 2 -b i )(w 2 -b 2 ), 
we form the two combinations 

(III.) y 2 Y' 2 + z 2 Z' 2 - ( Y 2 + Z 2 )R' 2 = 

(IV.) 2/ 2 Y' 2 -z 2 Z' 2 -(Y 2 -Z 2 )R' 2 = 



THE CURVE ON ONE OF THE COORDINATE PLANES. 479 

Remembering 



and that 



we get 



where 



y 2 z 2 = s, y 2 + z 2 = r 2 , 

2/V + z 2 y* — (y 2 + z 2 )s = r 2 s 
y 2 z* — z 2 y* = — (y 2 — z) 2 s , 

(III. ) ( r 2 X 2 + (b 2 - u 2 )(u 2 r 2 - 2¥)K 2 

0=S +2X2SXQXJ 
( + r 2 sX 1 2 

(IV.) ,X 2 j 

= (y 2 -z 2 ))-sX 1 2 C 

( +(6 2 -uVR' 2 ), 



X =u 2 [& 2 (Z 1 -w 2 )-<] 

u 4 

R' 2 = (t + L ) 2 - 8mu 6 s(i + L + L t ) . 
Let us separate the terms affected by s from those which are not. Then we get 

(III.) r 2 X 2 + (b 2 - u 2 )(r 2 u 2 - 2¥)(t + L ) 2 

+s|-4X X 1 +r 2 X 1 2 

L + (b 2 - u 2 )(u 2 r 2 - 2¥)( - 8m)u%t + L + Lj) 
(IV.) x 2 +u 2 (& 2 -u 2 )(* + L ) 2 



= 



^2 



+ s[ - X x 2 + u 2 (6 2 - u 2 )( - 8m)u 6 (i + L + L x )] 

this last neglecting the factor (y 2 — z 2 ), which gives a particular solution. 
Designating the two equations respectively by 

A+sB = (Ill') 

C-sD-0 (IV) 



and remembering 



& 



we 



have 



r 2 u 2 = t, -m = ^g- 4> 

A = fu 2 [6 2 (^-u 2 )-i] 2 
+ t(b 2 -u 2 )(L + t) 2 
-2¥(b 2 -u 2 )(L +t) 2 

B = t^(L 2 +t) 2 



+ (t - 2¥)(b 2 - ^){^){t + L + L X ) 



C = %*[&% - u 2 ) - tf + u 2 (b 2 - u 2 )(L + ty 
B = ^(L 2 +t) 2 -^(b 2 - U 2 )(L +L 1 + t), 

VOL. XXXV. PART II. (NO. 13). 4 K 



480 DR G. PLARR ON THE DETERMINATION OF 

We see that A and B are of the third degree in t, and C and D of the second degree. 

The elimination of s gives us 

BC+AD=0 (V) 

consequently of the fifth degree in t. We must remember that we have put 

(-D)=-X 1 Hu^-u^{+L 1 +L ) ; 

Let us state the coefficient of t 5 in the equation (V.), 
The coefficients of t z in A and B are respectively, 

mA, = tt 2 +(b 2 -u 2 = b 2 , 

inB ' = (^) Xl 
The coefficients of t 2 in C, D, are respectively, 

in C , = u 2 [u 2 + (b 2 - u 2 )] xl = u 2 b\ 

inD ' = (jp) xl - 
Therefore the coefficient of t 3 x t 2 or f in (V.) will be 

As the terms in A beside the term b 2 f are all divisible by b 2 , owing to L = b 2 (a 2 + u i ), 
and as the same remark applies to C in respect to bH 2 , and its other terms, we shall put 

A = b 2 A! 

C = u 2 b 2 C 
a*¥ 

Then the equations (III'), (IV) (by dividing the first by b 2 , the second by 6V) may 
be put under the forms, 

A' + ^B'=0 . . . (HI") 

C'-~D'=0 .... (IV") 

a 4 6 6 

and the s eliminated in 

B'C'+AT>' = 0, (V") 

where A', B', &c, are of the form, 

A=A +A 1 *+A 2 < 2 +* 8 

B=B +B 1 t + B 2 t 2 + t* 
G = C + C l t + t 2 
T> = J) + T> 1 t + t 2 . 



THE CURVE ON ONE OF THE COORDINATE PLANES. 
These coefficients A , A 1} &c, may now be deduced from the expressions, 

+ {t-2¥)(^^[b\a 2 +u 2 ) + tf . 

B'=t(L 2 +ty 

+ 4a 2 & 2 [6 2 (J 1 - u 2 ) - t](L 2 + 1) 
+ 8a\t-2b i )(b 2 -u 2 )(L + L x + 1) 

C'-jV(^-^ 2 )-^ + (^)(L + 2 

B' = (L. 2 + tf-8a 2 (b 2 -u 2 )(L + L 1 + t 2 ) , 

L = b 2 (a 2 +v 2 ) 
L!=u 2 (6u 2 -4y 

L. 2 = 2a 2 b 2 + L +L 1 

= 3a 2 6 2 - u 2 (4a 2 + 76 2 ) + 6u 4 . 



481 



where 



The factor 



su 6 _ y 2 z 2 u® 



is now of dimension zero. 

Considering s = y 2 z 2 in itself we may transform it in referring p to two axes /, ¥, of 




which / is the bissecting line of the angle between j, h 

Having then 

p=jy+1cz=j'y'+k'z\ 
we get 

y'~ z ' 1 , , . >\ 

y =J2~> Z =T2 iy+Z) ' 
and 

We put 

y'=r cosS 

z'=r sin (£ 



482 



DR G. PLARR ON THE DETERMINATION OF 



then 

and remembering r*u* = t, we have 

Thus 
and also 



s = - r r 4 cos 2 2(J, 



Sr« = £^ (cos2(E)2 



2onr A' 4a 4 6 6 



(III") 
(IV) 



The term independent of t and u 2 in (V") will be, 

~4>a?b*M v 3a 2 b 2 "1 x [a 4 6 4 ] 
_-8a 2 2& 4 .6 2 .a 2 & 2 _ 
+ [ - 2b* x 1 x a 4 6 4 ] [9a 4 & 4 - 8a 2 6 2 . a^ 2 ] 
12a 4 &%-16a 4 & 8 1a 4 & 4 xl 
_-2a 4 6 8 J 

= a 8 6 10 [12(a 2 + 26 2 ) - 166 2 - 26 2 ] 
= a 8 6 10 [12a 2 +66 2 ]. 

This shows that the equation (V") does not admit the factor t. 
We have promised to establish the identity between 

M_i M\ W\ 

(Xj Q>2 ^3 

where a x , a 2 , a s are the scalar coefficients of 

a = ia 1 +ja 2 + ka 3> 

and where M_ u M/, M/', are the expressions which we have quoted at the beginning 
We take R l5 R 2 under the form 



m \ wu/ 

\ nv) 



m 



Having 

Sj<f>jSk<j>Jc-S 2 j<j>k = 
as was shown before, we shall have 



■u z 



m 



+ 



Wi-^Vi-^). 

\ uv/\ wu/ 



THE CURVE ON ONE OF THE COORDINATE PLANES. 483 



Dividing by w 2 , and ordaining according to 

JL j_ _l 

UV ' WU ' VW ' 



we get 



U 4 Mi r i +w2 §M±s^ +u «n 

L m m J 





wu L m 



+ 
vw 

Let us bring w 4 M/ to the same stage. We get 

u^ = (uPjX^Q) - (wP 2 )(u 3 R 2 ) 



Thus 



(x x u 2 \Sj<f>k 
\ uv *v w- 

_/^_ xrs^ 4 /_^x-| 

\WU f\_ UV \ UV/J 

, u 2 r Sj<j>k 2 ,~| 

H cci-^ - — u 2 z, 2 I 

-u/uL w- J 



x x u 2 {Sk<f>k 

4 



— =— I — ■— +u- 

wu\ m 



') 



CuyZ^Uj 



vw 
We now recall (where n = a 2 — b 2 ) 

<f>~ 1 u> = anSaw — b 2 w 



These give 



also 



^ = b 2 d>- 1 o ) +w(a 2 +b 2 ). 

u 2 = b 2 + na^ \ (®i = na 2 a 3 
v 2 =b 2 +na 2 > > -J y x = na % a x 
w 2 = b 2 +na s 2 ' ^z 1 =na 1 a 2 , 

m Ji m 

+ Si * i + m2 =+L =+b 2 (a 2 +u 2 ) 



~- = a 2 ¥ 

m 



484 DR G. PLARR ON THE DETERMINATION OF 

Owing to 

Sj<j>j + Sk<pk _ S i<pi + m 2 
m m 

the term in M, in the first line will be 

a 2 fc 4 -<w, 2 6 2 (a 2 +u 2 )+u 6 

= u 4 (u 2 -6 2 )-a 2 6 2 (u 2 -& 2 ) 
= (u 4 — aW)na^ . 

We need not transform the other terms of M l5 except by changing the factors z x , y u y u z u 

respectively into 

na x a 2 , na^ , n 2 a^a 2 a z ; 

thus 

1 = (u 4 — a 2 ft 2 )na 1 



\ m ) v \ m / w 



vw 
Proceeding to transform w 4 M/, and considering its first term, we remark 

Also 

Sj^-HSk^k-Skcfi-HSjcp-ik , 

namely, 

Zl w* - y^ = SVjkY<p -ik$-H = ^ = b% , 
we get for the term in question, 

1 1_ on J 

But 

m 

Hence the term becomes 

^[y-a 2 ^]. 
The second term in w 4 M/ is 

The factor between [ 

= na^\b 2 na^ — u^.na^] 
= na 2 [b\w 2 - 6 2 ) - u\u 2 - b 2 )] 
= na*[b\{w 2 + u*) - V] - u 4 ] 
= na£\b\a? + b*-v i )-u i ] 



% J 



= —na 9 2 \ u i + -^ 
' |_ or, 

The third and fourth terms in M/ are respectively 

— na 2 a 3 



\ on /' 



w\ on 



THE CURVE ON ONE OF THE COOEDINATE PLANES. 



485 



and 
Hence 



lb Cv-i (A/a (X/n 

vw 



u*M,' 



a, 



- = the same expression than = . 

1 a, 



In the case of M/' we know that 



which, as 



^K = 



rJ,K= -ifM^ + M^Q +M" 1 R 2 ] 
+ fc[M 1 P 2 +M' 1 R 1 + M" 1 Q], 



gives zero identically for the coefficients of j and k We have then by the second of 
these identities 



m» n ( M i\r x) , t> i M'j Mj 



or 



M/'w 3 



M~t) 



\ m w J 



u° 



Now the coefficient of — is 



w 



a^ x — a 2 y 1 = zero. As — Oj X z x = — na^a 2 , 
we have 

The expression in [ ] is 



u\b 2 - w 2 ) + 6 2 (v 2 - a 2 - 6 2 ) + w 4 
= b 2 [u 2 + v 2 -a 2 -b 2 ] = b 2 (b 2 -w 2 ) 



We have then 

But 
hence 



= — b 2 na 2 . 



M /M\ 

M" x w 3 Q = — I . a 2 b 2 na 3 2 = ( — l )a 3 & 2 a;i . 

(Xj \ a l/ 



% 3 Q=6 2 a;i ; 
a 3 — a x ' 



Thus in the case b 2 = c 2 , the three expressions of Mi, M/, M/', when equated to zero, 
give only one distinct equation. 

We may remark that M/ and M/' may be annulled by putting 

P 1 = 0,P 2 = 0. 

These two equations give four solutions as to a 1} a 2 , a 3 , of which one is also given by 



a x 2 = 


a 2 = 


2 1 

a 3 - 3 


<^l 2 + «2 2 + a 3 2 


= !)• 


u 2 = v 2 


= w 2 


-gfcl, 



486 DR G. PLARR ON THE DETERMINATION OF 

3l, = , although Mi is expressed independently of P lt P 2 . The equations Pi = , P., = 
give 

ux 1 =vy 1 = wz 1 . 

Squaring we get by the two first members 

(b 2 + na x 2 + )n 2 a 2 2 a 3 2 — (b 2 + na^)n 2 a 3 2 a^ = . 
This gives, with the third member wz x treated in the same way, 



(owing to 

Consequently 



and u — v = w, as all three must be positive, and u 2 + v 2 + w 2 = l x = (a 2 + 2b 2 ). 

Let us substitute these elements into M : and verify the signs which can be attributed 
to the first powers of a u a 2 , a 3 ? 

We have now owing to u 2 = v 2 = w 2 

m m 

Also 

u _ u _ u2 _ i 

V w vw 

= na-> (% 4 — a 2 b 2 ) 

a x x ' 

— {na 2 + 7ia 3 )[# + b 2 u 2 — a 2 b 2 — 6 4 ] 
+ n 2 u 2 a x a 2 a 3 . 

In the second term grouping u^ — a 2 b 2 together, and replacing b\u 2 - h 2 ) by nb 2 a 2 , we get 

— — - — n\a x — a 2 — a 3 ](« 4 — a 2 b 2 ) 

+ n 2 [u 2 a 1 a. 2 a 3 — b' 2 a 1 2 (a 2 +a 3 )] . 
We have 

M 4 _ a 2J2 = ( a 2 ai 4 _ ^2 a '4) X n ( 

where a' 2 = 1 —a 2 . (This transformation can be shown by considering 

u 2 = b 2 + na 2 , and u 2 = a 2 — n 2 a' 2 . ) 

We divide by n 2 , and replace u 2 by u 2 = a 2 a 2 + b 2 a' 2 ; then we get 

w 4 M 

_i = («j _ a 2 - a 3 )(aV - 6V 4 ) 

+ [a 1 a 2 a 3 (a 2 a 2 1 + & 2 a' z ) — b 2 a x \a 2 + a 2 )] 
= a 2 [a 1 4 (a 1 — a 2 — a 3 ) + a! 3 a 2 a 3 ] 

+ b 2 r~a' 4 (a 2 + a 3 — aj) — aj 2 (a 2 + a 3 )" 



' pa' 4 (a 2 + a 3 — a x ) — aj 2 (a 2 + a 3 )~j 

L + ft' 2 (ai«2 a 3 J 

= c^a/Kc^ — a 2 )(ai — a 3 ) 



THE CURVE ON ONE OF THE COORDINATE PLANES. 



487 



+ b 2 



4 1 

g(«2 + <V- a l)— 3O2 + «s) 



(because of a 2 = ^, a' 2 = § ). Hence 



— ^-i = « 2 a 1 3 (a 1 — a 2 X a i — a 3 ) 



17, Ct, 



6 2 
+ g[(a 2 + %) + Sa^aaaffg - 2)] . 

If now we had a x = a 2 — —a 3 , then (as they are all three of the same absolute value), the 
coefficient of a 2 would vanish, but that of b 2 would be negative 

= ^(-3^-2) 

and Mj could not vanish. 

If a x = — a. 2 = —a 3 , then the coefficient of b 2 would become 

^2a 1 [-l + 3a 1 2 -2] = ^| ) 

4 
whereas the coefficient of a 2 would become + - , so that M x would vanish only for a 2 — b'\ 

J v3 

There exists therefore the only solution 

j_ l l 
a 1 = a 2 = a 3 =-\ — -=, or = -=, 

which annuls M 1( namely, the direction for a of the median line of the octant i, j, k (or 
its opposite direction, which does not alter the circumstances), 

The values of y and z now become equal, that is 

— - n\ 

1 =»&-»)]. 



As 



This gives 



l 1 -n = a?+2b 2 -(a 2 - b 2 ) = 3& 2 
&V3 

2 2 , 2 6&4 



and 

t = ti 2 r 2 = 2b*. 

We have thus a system of values of t, u 2 , namely 2& 4 and \l x respectively, which solve 

the equation (V"). It is, however, a certainty that this point y = z = — ^=. does not 

v n 

belong to the outward branch of the limiting curve, but that it forms a singular point of 
a branch (or branches) contained in the inside of the limiting curve as defined by \\ie = 0. 

VOL. XXXV. PART II. (NO. 13). 4 L 



( 489 ) 



XIV. — On Ostracoda collected by H. B. Brady, Esq., LL.D., F.R.S., in the South Sea 
Islands. By George Stewardson Brady, M.D., LL.D., F.R.S. (Plates I.-IV.) 

(Read 3rd December 1888.) 

Excepting the few species noticed in the Report on the Ostracoda of the "Challenger" 
Expedition, scarcely anything, so far as I know, has been published respecting the 
Ostracoda of the South Sea Islands. Prof. G. M. Thomson has indeed published in the 
Transactions of the New Zealand Institute (1878), a paper on Crustacea, which includes 
a few marine and fresh-water Ostracoda of New Zealand ; and the Rev. R. L. King, in the 
Proceedings of the Royal Society of Van Diemens Land (1855), described numerous 
species of Entomostraca, amongst which were several fresh-water, but no marine, 
Ostracoda. Dr Baird also published a species of Cypridina from New Zealand. I have 
myself contributed to the Proceedings of the Zoological Society of London (1886) a 
paper on Entomostraca collected in South Australia, chiefly by Professor Ralph Tate of 
Adelaide, including a considerable number of fresh- water Ostracoda ; and in a French 
publication (Les Fonds de la Mer), edited by the Marquis de Folin, there are likewise, by 
myself, descriptions of a few species taken at Noumea, New Caledonia. There are also, in a 
paper of mine published in the Transactions of the Zoological Society (1865), notes of 
a few Australian marine species. This, I think, represents the sum of our present know- 
ledge respecting the Ostracoda of these regions. 

The collection to be noticed in this memoir was taken entirely from material obtained 
either between tide-marks, or from very small depths of water — not as a rule exceeding 
6 fathoms. The material so obtained was, however, not collected with any view to the 
Ostracoda, and having been preserved in a dry condition, it has been impossible to obtain 
details of internal structure, as might have been done with spirit preparations. Besides 
the fact of a large proportion of the species being new to science, the collection presents 
the following points of interest: — First, some species, notably Bairdia amygdaloides and 
Bairdia foveolata, were found in considerable numbers, in fine condition and in various 
stages of growth, so that I have been able better to define and emphasise their characters, 
and so to place those species on a more stable foundation. It would have been interest- 
ing, had space permitted, to have given a series of drawings representing stages of growth 
and other variations in those species, but the more important points will be found briefly 
noticed in the text. Secondly, it would seem, from their abundance in some of these 
gatherings, that various Cypridinidse occur in the living condition in great numbers 
between tide-marks. I am not aware that in the Northern Hemisphere any member of 
this family has ever been taken * except by the dredge, or in the tow-net over deep 
water. Professor G. M. Thomson, however, mentions a species (Philomedes agilis) as 

* Except once in Herm, by the Rev. Dr Norman. 
VOL. XXXV. PART II. (NO. 1 4). 4 M 



490 DR G. S. BRADY ON 

occurring in rock-pools in New Zealand. The genus Sarsiella (a Cypridinid) is strongly 
represented in tidal pools, as is also a new genus Pleoschisma, which is closely allied to 
Cypridina. On the whole, it would appear that an investigation of the littoral zone of 
these islands would acquaint us with forms of the highest interest belonging to this 
particular group. The collection contains no gatherings from fresh water, but there 
occur several shells which apparently belong to fresh-water genera. These, as they are 
found only singly, I have not described. But it is interesting to note that there is one 
undoubted example of Cypris obliqua, Brady, a well-known British species, and another 
which perhaps may belong to Cypria ophthahnica, Jurine (compressa, Baird). One speci- 
men belongs apparently to the genus Limnicy 'there. This, as it presents sufficiently 
distinct characters, I have described and figured. Besides these there are a few specimens 
belonging probably to Cypris or Candona. All of them are probably interlopers, washed 
down from fresh water. I have not thought it necessary to insert a complete bibliography 
of the species, but have given references in all cases where they occur in the " Challenger" 
Report. 



Section I. PODOCOPA. 

Fani. Cypridid^;. 

Genus Phlyctenophora, G. S. Brady. 

Phlyctenophora viridis, n. sp. (PI. I. figs. 1, 2). 

Shell, seen from the side, elongated, subtriangular, highest just in front of the middle, 
height equal to half the length ; dorsal margin obtusely angulated at its highest point, 
thence sloping almost in a straight line towards the front and with a gentle curve to 
the posterior extremity, which is placed altogether below the middle of the valve ; ventral 
margin almost straight; anterior extremity broadly and evenly rounded, posterior narrow 
and rounded. Seen from above, the shell is ovate, widest in the middle, not quite thrice 
as long as broad, the sides rather boldly curved ; extremities subacute, the posterior 
somewhat the more compressed of the two. Valves thin, pellucid, smooth and polished, 
greenish, variously mottled with patches of a darker hue, and showing faint longitudinal 
striations after the manner of Cypria exsculpta. Length "80 mm. 

The verticiUate sac ("glandula mucosa"), antennas, and feet agree generally with 
those of the type Phlyctenophora zealandica., described in the Report of the " Challenger " 
Expedition. The caudal rami are slender and destitute of marginal setae, the apices 
bearing a single small seta and two long curved claws which, on the concavity close to 
the apex, have two or three lateral cilia. This pretty species appears to be one of the 
commonest and most characteristic littoral species of Samoa and Fiji, occurring plentifully 
in almost all the gatherings from those groups, as well as in dredgings from the Port of 
Noumea. The best preserved and most highly coloured specimens are those from tidal 
pools ; dredged specimens are usually only empty shells. 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 491 

Phlyctenophora (?) reniformis, n. sp. (PL I. figs. 9, 10). 

Shell, seen from the side, subreniform, greatest height situated in the middle and equal 
to half the length. The anterior extremity is broad and well rounded, the posterior 
rather narrower ; dorsal margin well arched, sloping gently to the front and with a steep 
curve backwards ; ventral margin very slightly sinuated. Seen from above, the outline 
is compressed, ovate, widest in the middle, about three times as long as broad; anterior 
extremity acutely, posterior subacutely pointed. Shell smooth, almost colourless, slightly 
mottled. Length *75 mm. 

A few empty shells of this species occurred in shore-sand at Loma-Loma, in material 
from Suva mud-flats, and from the reef at Lufi-Lufi ; but the specimens have probably 
lost their original colour, and the generic reference is doubtful. 



Genus Pontocypris, G. 0. Sars. 
Pontocypris attenuata, G. S. Brady (Plate I. figs. 3, 4). 

Pontocypris attenuata, Brady, Ann. and Mag. Nat. Hist., ser. 4, vol. ii. p. 179, pi. iv. 6gs. 11-14. 
? Pontocyj/ris nitida, Brady, Linn. Soc. Jour. (Zoology), vol. xix. p. 303, pi. xxxix. figs. 4-6. 

The specimens from which this species was originally described differ from those here 
figured both in shape and size, being considerably smaller, not so high in proportion to 
their length, and having no posterior spine. The full-grown South Sea specimens have 
a height more than equal to half the length, and are armed at the postero-ventral angle 
with a single short but stout spine. The surface is very faintly punctate, and is densely 
clothed with fine hairs. The types were taken at Mauritius, and the collections described 
in the present memoir contain young specimens of exactly the same character. The 
species occurred in dredgings from the port of Noumea and in shore gatherings from the 
reef at Apia, Upolu. The length of the adult is 1*075 mm. 

A Ceylon species, described by me in the Journal of the Linnean Society as 
Pontocypris nitida, may perhaps belong to P. attenuata, and is probably the very 
young form of that species. 

Pontocypris gracilis, n. sp. (PL I. figs. 5,6). 

Shell, seen from the side, siliquose, much elongated, greatest height in the middle 
and equal to less than one-third of the length, depressed and rounded in front ; posterior 
extremity much tapered and subacute, scarcely rounded. Dorsal margin arched, sloping 
with a gentle curve to the front and much more steeply behind; ventral margin almost 
straight. Seen from above, elongated, ovate, about four times as long as broad, widest 
near the middle, extremities acute. Surface of the valves quite smooth, bearing a few 
very small, distant papillse. Length TO 7 mm. 

Habitat. — Between tide-marks, Levuka and Eambe Island. 



492 DR G. S. BRADY ON 

Pontocypris sicula, n. sp. (PI. I. figs. 7, 8). 

Shell, seen from the side, slender, awl-shaped; greatest height equal to about one-third 
of the length, and situated near the middle. Anterior extremity depressed, rounded and 
narrowed ; posterior excessively depressed, produced and tapered to an acute point on the 
level of the ventral border ; dorsal margin arched, highest near the middle, forming a 
gentle curve in front, but sloping steeply and in a right line quite to the posterior 
extremity ; ventral margin straight. Seen from above, the outline is lanceolate, four 
times as long as broad, broadest in front of the middle, acutely pointed behind, subacutely 
in front. Surface of the valves smooth, covered with closely-set, minute, impressed 
puncta. Length 9 mm. 

Habitat. — Sava-Sava Bay, 4 fathoms. 

Fam. Bairdiid^e. 
Genus Macrocypris, G. S. Brady. 

Macrocypris decora, G. S. Brady. 

Macrocypris decora, Brady, " Ostracoda of Challenger Expedition," p. 44, pi. i. fig. 3 a-d and 
pi. vi. fig. 8 a-b. 

Found in shore-sand, Porcheron's Beach and near Artillery Point, Noumea ; dredged 
in the port of Noumea, 3-6 fathoms ; and between tide-marks, Vuna Point, Taviuni. 

Genus Bairdia, M'Coy. 

Bairdia simplex, G. S. Brady. 

Bairdia simplex, Brady, " Ostracoda of Challenger Expedition," p. 51, pi. vii. fig. 1 a-d. 

Habitat. — Vuna Point, Taviuni, Fiji, between tide-marks. The types were taken by 
the " Challenger" off Heard Island. 

Bairdia amygdaloides, G. S. Brady. 

Bairdia amyydaloides, Brady, " Ostracoda of Challenger Expedition," p. 54, pi. ix. fig. 5 a-f, 
pi. x. fig. 2 a-e. 

Habitat. — Port of Noumea, 3-6 fathoms ; Suva, reef ; Levuka, between tide-marks ; 
Mango Island, fringing reef; Rambe' Island, between tide marks ; amongst shore-sand, 
Loma-Loma ; Vuna Point, Taviuni, between tide-marks. 

A very fine series of this species occurred in several of the gatherings above mentioned. 
In the living condition the shell is beautifully blotched with chocolate-brown, and is 
usually smooth or nearly so — never more than very moderately punctate, sometimes 
also bearing a few scattered silky hairs. The posterior extremity never has any well- 
developed beak, but ends acutely, sometimes in a single small spine. The marginal 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 493 

serratures are very variable, but the anterior extremity is often minutely serrated, and 
the posterior portion of the ventral margin generally has a series of somewhat larger 
teeth. Seen dorsally, the outline presents considerable variation, being sometimes wide 
in the middle and very boldly and evenly curved ; in other cases the greatest width is in 
front of the middle, and the posterior portion is tapered. I have not been able to make 
out the meaning of these differences. In such specimens as I have dissected the 
rounded, tumid shells were, contrary to my anticipation, males. The more tapered 
form, with anterior tumidity, may perhaps be the female, but of this I am not sure. 
The figures given in the "Challenger" Report represent tolerably well the centrally 
tumid form only. 

Bairdia tenera, G. S. Brady (PI. I. figs. 11, 12). 

Bairdia tenera, Brady, " Entomostraca collected in Ceylon," Jour, of Linn. Soe. (Zoology), 
vol. xix. p. 304, pi. xxxix. figs. 13-15. 

Shell, seen from the side, subreniform, highest in the middle, height somewhat less 
than two-thirds of the length; anterior extremity wide, obscurely angulated at its 
junction with the dorsal margin, and obliquely rounded below ; posterior extremity some- 
what produced in the middle, but scarcely beaked, obliquely rounded off below ; dorsal 
margin nearly flat for the greater part of its course, and sloping abruptly towards both 
extremities ; ventral distinctly sinuated in the middle, and minutely dentated towards 
the posterior extremity. Seen from above, compressed, ovate, widest in the middle, not 
quite thrice as long as broad ; extremities obtusely pointed. Surface of the shell smooth, 
beset with numerous very fine but rigid short hairs, and marked with closely-set im- 
pressed puncta. Length '85 mm. 

Habitat. — Pools on reef, Lufi-Lufi ; and Apia, Upolu.* 

Bairdia Crosskeiana, G. S. Brady. 

Bairdia Crosskeiana, Brady, " Ostracoda of Challenger Expedition," p. 58, pi. ix. fig. 3 a-e. 
Habitat. — Mud-flats, Suva ; pools on reef, Lufi-Lufi ; and Apia, Upolu. 

Bairdia foveolata, G. S. Brady. 

Bairdia foveolata, Brady, "Ostracoda of Challenger Expedition," p. 55, pi. viii. fig. 1 a-f, and 
fig. 2 a-f. 

Habitat. — Noumea, shore-sand, and dredged in 2-6 fathoms ; mud-flats, Suva ; Sava- 
Sava Bay, 4 fathoms ; Lufi-Lufi, Samoa, pools on reef ; Apia, Upolu, tidal pools ; Mango 
Island, Fiji, fringing reef ; Vuna Point, Taviuni, between tide-marks. 

This species is largely represented in many localities. Like B. amygdaloides, it is in 
the living condition beautifully marked with blotches of a chocolate-brown colour, and is 

* The single specimen from which the species was first described was unfortunately lost by the draughtsman who 
made the drawings, but these Samoan shells seem to agree almost exactly with the description, and I have but little 
hesitation in referring them to the same species. 



494 DR G. S. BRADY ON 

always covered with closely-set impressed punctations. These are even more strongly 
developed in very young than in older shells, where they tend to become obliterated by 
calcareous deposit. The beak is generally distinct, but never very largely developed. 
The anterior margin, and the posterior margin below the beak, are beset with short, 
blunt teeth, which are usually irregular, as if broken away in places. The specimens 
figured in the " Challenger" Report are probably rightly referred to this species, but being 
all dredged in deep water, were old and worn shells deficient in surface ornament. Some 
specimens of B.foveolata approach very closely B. Milne- Edwardsii, and I am not sure 
that further research may not show that both forms belong to the same species. 

Bairdia Milne- Ed wardsii, G. S. Brady. 

Bairdia Milne-Edwardsii, Brady, " Ostracoda of Challenger Expedition," p. 56, pi. x. fig. 4 a-g. 
Habitat. — Noumea, 2-4 fathoms; Levuka, between tide-marks, Loma-Loma, shore- 
sand ; reef and shore-pools, Apia, Upolu ; Suva, pools inside reef. 

Bairdia ventricosa, n. sp. (PI. IV. figs. 17, 18). 

Shell, seen from the side, oblong, subrhomboidal, nearly equal in height throughout, 
height equal to half the length. Anterior extremity rounded and very finely serrated 
below the middle, posterior sloping very steeply and in an almost straight line below the 
middle, then abruptly rounded off to the ventral margin, edge serrated ; dorsal margin 
straight, ventral also straight and parallel with the dorsal. Seen from above, oblong, 
subovate, with moderately arcuate sides and produced mucronate extremities, posterior 
extremity more compressed than the anterior ; greatest width in the middle and equal to 
half the length. Shell-surface marked throughout with closely-set, small, circular impres- 
sions. Length '75 mm. 

Found in shore-sand from low- water, near Artillery Point, Noumea. 



Bairdia hirsuta (?) G. S. Brady. 

Bairdia hirsuta, Brady, " Ostracoda of Challenger Expedition," p. 50, pi. viii. fig. 3 a-d. 
Habitat. — Port of Noumea, 3-4 fathoms. 

Bairdia Woodivardiana, G. S. Brady. 

Bairdia Woodwardiana, Brady, " Ostracoda of Challenger Expedition," p. 57, pi. xi. fig. 1 a-e. 
Habitat. — Vuna Point, Taviuni, between tide-marks. 

Two empty shells of this very curious species were found. They agree very closely 
with those figured in the " Challenger " Report, but are even more attenuated behind, and 
one of them is decidedly more obtuse in front on the dorsal view. 

Bairdia truncata, n. sp. (PI. II. figs. 1, 2). 

Shell, seen from the side, oblong, irregularly angular, height equal to half the length. 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 495 

Anterior extremity narrow, scarcely rounded, bearing three or more small blunt, forward- 
pointing spines ; posterior obliquely truncated, narrow, lying altogether below the middle 
line, the obliquity looking downwards, divided into about six strong, sharp teeth ; dorsal 
margin straight in the middle, sloping abruptly and almost in a right line at each 
extremity ; ventral margin straight. Seen from above the outline is very irregular, 
oblong-ovate, twice- as long as broad, widest in the middle, extremities wide, truncated, 
and spinous. Shell-surface marked everywhere with rather large, irregular, subangular 
impressed puncta ; abruptly depressed within the anterior margin and over the posterior 
extremity. Length "56 mm. 

Habitat. — One perfect specimen of this curious species was found in a gathering 
from pools on the inner reef at Apia, Upolu ; another single valve in shore-sand from 
Porcheron's Beach, Noumea. 

The uneven outline of the shell, as seen dorsally, seems to indicate a distortion or 
malformation; but, apart from this, the characters are so peculiar that there can be no 
doubt as to its specific distinctness. 

Bairdia nodulifera, n. sp. (PL I. figs. 13-16). 

Shell, seen from the side, subreniform, highest in the middle, height equal to more 
than half the length. Anterior extremity obliquely rounded and minutely serrated below 
the middle, posterior produced below the middle into an obtusely angular beak, above 
which it slopes forwards with a distinct sinuation, the curve encroached upon by two slight 
rounded projections and minutely serrated below the middle ; dorsal margin boldly arched, 
ventral slightly sinuated in the middle. Seen from above, compressed, ovate, widest in the 
middle, more than twice as long as broad; lateral margins evenly and moderately convex, 
twice or thrice emarginate near the extremities ; extremities obtuse, subtruncate. Surface 
of the valves smooth, marked with closely set, small, circular impressions ; within the 
anterior border are three or four large, but not very prominent, rounded tubercles, and 
the posterior margins have a somewhat similar armature. These tubercles are not at all 
conspicuous when seen laterally, but the marginal irregularities produced by them, when 
seen dorsally or ventrally, are very characteristic. Right valve smaller and more angular 
than the left. Length - 8 mm. 

Habitat. — Levuka, between tide-marks. One specimen only seen. 

Bairdia tuberculata, G. S. Brady. 

Bairdia tuberculata, Brady, " Ostracoda of Challenger Expedition," p. 60, pi. x. fig. 3 a-d. 
Habitat. — Port of Noumea, 3-6 fathoms. 

Bairdia expansa, G. S. Brady. 

Bairdia expansa, Brady, " Ostracoda of Challenger Expedition," p. 58, pi. xi. fig. 2 a-e. 

Habitat. — Lufi-Lufi, reef and shore-pools ; Apia, Upolu, reef and shore-pools 
Noumea, between He Pore-Epic and shore, 2-6 fathoms. 



490 



DR G. S. BRADY ON 



Genus Anchistrocheles, Brady and Norman.* 

Shell reniform, mucli compressed ; anterior extremity very oblique, ventral margin 
deeply sinuated. Antennules six-jointed,tthe last five joints nearly equal and very short, 
their united length only equalling about one-third of the first joint, hairless except the first 
and fifth joints, each of which bears a single seta of moderate length ; the only fully setiferous 
joint is the last, which bears an apical brush of about ten long setae ; antennae five-jointed, 
bearing at the apex two rather long setae and a still longer curved claw, which consider- 




Anchistrocheles furmta. 

1, Antennule ; 2, Antenna ; 3, Mandible ; 4, Maxilla ; 5, Branchial plate 
of maxilla ; 6, Foot of first pair (a) rudimentary second maxilla ; 
7, Foot of second pair ; 8, Caudal ramus ; 9, Copulative organ of 
male (v.d.), vas deferens (all much magnified). 

ably exceeds in length the entire limb, and is bent rectangularly at the apex, so as to form 
a minute hook. Mandible slender, toothed at the apex ; first joint of the palp bearing a 
small trisetose branchial appendage, but otherwise almost destitute of setae, except at the 
apex. First pair of maxillae provided with a branchial plate of moderate size ; second 
pair (?) rudimentary, in the form of a setiferous, one-jointed appendage attached to the 

* Trans. Royal Dublin Soc, vol. iv. Series II. p. 110. 

t These anatomical details being taken from dried specimens, were not easily made out, and the number of joints 
here given differs from that of the Royal Dublin Society Memoir. I think the present enumeration is correct. 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 497 

basal joint of the first foot. Two pairs of feet, both of them clawed and adapted for 
walking, the claw of the first pair hooked (as in the antennae). Caudal rami extremely 
small, bearing three apical setae, two of them long and one very minute. Copulative 
organs of the male large and complex. 

Anchistrocheles fumata, n. sp. (PI. III. figs. 13-14). 

Shell thin and fragile; seen from the side, reniform, greatest height equal to about half 
the length and situated in the middle ; anterior extremity only slightly rounded, wide, 
obliquely subtruncate, the obliquity looking downwards and forwards ; posterior narrowed, 
somewhat produced in the middle, and rounded ; dorsal margin very gently and evenly 
arched, ventral deeply incurved in the middle. Seen from above the outline is elongated, 
ovate, about four times as long as broad, with nearly parallel sides ; anterior extremity 
acute, posterior rounded. Shell smooth, transparent, smoky yellow with darker clouded 
patches. Length '75 mm. 

This very interesting species was found only in one gathering from shore-pools at 
Lufi-Lufi, Samoa. It forms a connecting link between the typical Cyprididee and 
Cytheridse, the antennules having the long setose lash of a true Cyprid, while the antennas 
possess only a few rudimentary hairs in place of the usual fascicle of setae ; the abortive 
character of the second maxilla shows an approach to the Cytheridse, while the presence 
of only two pairs of feet — the second of which, however, has the character belonging to 
that structure in the Cytheridse — indicates another approximation to the Cyprididae. 
The British species, A. acerosa is known from the shell only, and, until the discovery of 
this Samoan species, was provisionally placed in the genus Cythere. 

Fam. Cytheridse. 
Genus Cythere, Muller. 
Cythere demissa, G. S. Brady. 

Cythere demissa, Brady, " Ostracoda of Challenger Expedition," p. 66, pi. xii. fig. 7 a-j. 

The specimens here referred to agree closely with those figured in the " Challenger " 
Report. The species seems to be common and widely distributed, but varies a good deal, 
especially in the number and development of the teeth on the posterior extremity. 

Habitat. — Noumea, in shore-sand, and dredged in 2-6 fathoms ; Levuka, between 
tide-marks; Sava-Sava Bay, Vanua Levu, 4 fathoms ; Mango Island, fringing reef ; Rambe 
Island, between tide-marks ; Lufi-Lufi, Upolu, shore-pools and reef between tide-marks. 

Cythere crenata, n. sp. (PI. II. figs. 35-36). 

Cythere crispata, Brady, " Ostracoda of Challenger Expedition," p. 72, pi. xiv. fig. 8 a-d. 
Shell, seen from the side, oblong, subreniform, highest near the front, height equal to 
at least half the length. Anterior extremity broad and well rounded, posterior narrowed, 

VOL. XXXV. PAET II. (NO. 14). 4 N 



498 DR G. S. BRADY ON 

truncated, rounded off below, but forming an obtuse angle at its junction with the dorsal 
margin. Dorsal margin prominent over the anterior hinge, thence sloping with a slight 
outward curve to the posterior extremity, in front of which it is slightly emarginate ; 
ventral margin gently sinuated in the middle. Seen from above, ovate, more than twice 
as long as broad, circumference irregularly crenated or emarginate, sides subparallel, 
extremities obtusely rounded, but not at all truncated, the posterior considerably the wider 
of the two. Surface marked with undulated ridges very variable in their development 
(but towards the posterior extremity disposed more or less transversely), and enclosing 
between them fossae of irregular shape and size. Length '5 mm. 

Habitat. — Noumea, shore-sand, also dredged in 2-6 fathoms ; Suva reef, Levuka, 
between tide-marks , Sava-Sava-Bay, Vanua Levu, 4 fathoms ; Eambe Island, between 
tide marks ; Apia, Upolu, reef and shore-pools ; Mango Island, on fringing reef. 

The specimens referred in the " Challenger " Eeport to Cythere crispata are undoubtedly 
identical with those here described, and bear certainly a very close resemblance to the 
European species Cythere crispata, but after examining minutely the large series of 
specimens found in these gatherings from Fiji and New Caledonia, I no longer think that 
they are properly referable to that species. The outline of the northern form is more 
flexuous, the ribs generally fewer and more prominent and enclosing larger hollows, and 
viewed dorsally the posterior extremity is distinctly truncated. Individual specimens, 
however, vary very greatly, and it is by no means easy to assign distinct limits to the 
three species crispata, crenata and canaliculata. 

Cythere ochracea, n. sp. (PL II. figs. 8, 9). 

Shell, seen from the side, elongated, subsigmoid, greatest height in the middle, and 
equal to less than half the length. Anterior extremity evenly rounded, posterior rounded off 
below, and obscurely angulated above at its dorsal termination ; dorsal margin gently and 
evenly arcuate, ventral incurved for the greater part of its length, slightly convex behind. 
Seen from above, the outline is elongated, hastate, widest near the posterior extremity, 
width and height equal ; from the widest point the sides approximate with a gentle curve 
to the front, which is attenuated and acute, while behind they converge suddenly with a 
slight sinuation, terminating in a subacute, slightly produced angle. Surface rather coarsely 
pitted, the hollows running together so as to form flexuous grooves, which run mostly in 
an oblique direction, except on the anterior portion of the valve, where they are more or 
less distinctly concentric. In young adults the pitting is the most conspicuous feature, 
but in older specimens this gives place largely to the grooved sculpture. Colour 
yellowish-brown. Length *5 mm. 

Habitat. — Noumea; in shore-sand from Porcheron's Beach, and Artillery Point. 

Cythere injiata, n. sp. (PL II. figs. 3-5). 

Shell, seen from the side, oblong, slightly flexed, greatest height situated near the 
front, and equal to half the length. Anterior extremity rounded, produced slightly below 



OSTKACODA COLLECTED IN THE SOUTH SEA ISLANDS. 499 

the level of the ventral margin, posterior obliquely truncated, the obliquity looking down- 
wards, slightly convex and angulated both above and below ; dorsal margin highest in 
front, flattened in the middle, then sloping very gently to the posterior extremity, just 
in front of which it shows a short sinuation ; ventral margin slightly sinuated. Seen 
from above, the outline is wedge-shaped, the greatest width equal to half the length and 
situated near the posterior extremity, which is flattened, almost truncated, but widely 
rounded at the angles ; from the widest point the sides converge almost in a straight line 
forwards, showing a slight constriction just in front of the middle, and meeting rather 
abruptly at the extremity. Shell-surface marked throughout with rather coarse im- 
pressed puncta, and raised at the posterior extremity, so as to form a transverse ridge or 
crest, which, in the dorsal view, is seen as a prominent angle. Only in some specimens, 
however, is this ridge seen ; the figured specimen does not show it in the dorsal view. 
Length '65 mm. 

Habitat. — Suva, mud-flats between tide-marks ; Levuka and Vuna Point, Taviuni, 
between tide-marks ; Sava-Sava Bay, 4 fathoms ; Rambe" Island, between tide-marks ; 
Loma-Loma, shore-sand ; Lufi-Lufi, Upolu, shore-sand. 

Cythere ovalis, G. S. Brady (Plate II. fig. 12). 

Cythere ovalis, Brady, " Ostracoda of Challenger Expedition," p. 66, pi. xiv. fig. 4 a-d. 
Habitat. — Mango Island, 2-3 fathoms. 

Cythere caudata, n. sp. (PI. II. figs. 10, 11). 

Shell, seen from the side, oblong, subquadrate, more than twice as long as high. 
Anterior extremity evenly rounded, posterior produced, forming a wedge-shaped pro- 
jection, the apex of which is rounded off and situated below the middle line of the valve ; 
dorsal and ventral margins nearly straight and parallel, the ventral very slightly sinuated. 
Seen from above, oblong, subovate, three times as long as broad, sides parallel for the 
greater part of their course, gradually converging from the anterior third to the subacute 
extremity, dipping abruptly behind and terminating in a strong central mucro. Shell 
surface marked with closely-set, small, oblong depressions, the long diameters of which 
coincide with the long axis of the shell. Length - 46 mm. 

Habitat. — Sava-Sava Bay, Vanua Levu, 4 fathoms. 

Cythere Scotti, n. sp. (PL III. figs. 3, 4). 

Shell, seen from the side, subquadrate, higher in front than behind ; height equal to 
rather more than half the length. Anterior extremity wide and boldly rounded, and 
bearing a variable number of small fringing spines ; posterior imperfectly rounded, its 
lower half slightly produced and bearing about six blunt marginal teeth ; dorsal margin 
prominent over the anterior hinge, thence sloping gently backwards with a slight sinua- 
tion, and rounded at its junction with the posterior extremity, ventral almost perfectly 



500 DR G. S. BRADY ON 

straight. Seen from above, very broadly ovate, greatest width behind the middle, and 
equal to two-thirds of the length, lateral margins boldly curved and forming a continuous 
sweep with the wide anterior extremity ; from the widest part of the shell the sides 
converge abruptly backwards, and terminate in a wide, truncated, central prominence ; 
margins everywhere jagged or crenated. Shell-surface marked with a raised reticulated 
pattern, enclosing irregular polygonal excavations. Length 1 mm. 

Habitat. — Banc de l'Aiguille, Noumea, 2-4 fathoms. A very fine and well-marked 
species, which I have pleasure in naming after Mr T. Scott of the Scottish Fishery Board 
— a most able and industrious investigator of the Entomostraca and other Invertebrata. 

Cythere cuneolus, n. sp. (PI. II. figs. 6, 7). 

Shell, seen from the side, oblong, quadrangular, highest in front, height equal to fully 
half the length. Anterior extremity well rounded, posterior narrow, obscurely angulated 
above, rounded off below, irregularly fimbriated or dentated ; dorsal margin sloping with 
a slight convexity backwards, ventral sinuated near the front and gently up-curved towards 
the posterior extremity. Seen from above, the outline is hastate, with irregular margins; 
widest near the posterior extremity ; width and height equal ; from the widest point the 
sides converge with an abrupt slant, broken by a conspicuous intervening prominence, to 
the posterior extremity, which is wide and truncated ; towards the front, which is also 
wide, but rounded, the sides converge more gradually, but in an undulated line. Shell- 
surface covered with closely-set fossae, and showing also one or two very vaguely-marked 
flexuous ribs. Length "45 mm. 

Habitat. — Banc de l'Aiguille, Noumea, 2-3 fathoms ; Mango Island, on fringing reef ; 
and from shore-sand at Loma-Loma. This may perhaps prove to be only the young form 
of some other species, the few specimens seen varying considerably in character, — but I 
am not acquainted with any to which it can properly be referred. 

Cythere torticollis, n. sp. (PI. III. figs. 1, 2). 

Shell, seen from the side, oblong, somewhat higher in front than behind, height scarcely 
equal to half the length. Anterior extremity obliquely rounded, and slightly crenated 
below the middle, posterior produced in the middle, its lower half divided into five or six 
short and broad teeth ; dorsal margin prominent over the anterior hinge, thence sloping 
gently in a sinuous line backwards ; ventral margin slightly sinuous throughout. Seen 
from above, the outline is very irregular, twice as long as broad, widest near the posterior 
extremity ; lateral margins deeply indented at several points, converging suddenly and 
in a very zig-zag line to the posterior extremity, which forms a triangular, centrally- 
emarginate prominence ; the anterior extremity is wider, rounded, and emarginatc in 
the middle. Surface of the valves very rugose, a wide rib just within and parallel to 
the anterior margin, an abrupt and irregular transverse ridge near the posterior extremity, 
and irregularly disposed ridges with numerous large fossse on the central portion. Length 
5*8 mm. 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 501 

Habitat. — Noumea, in shore-sand from near Artillery Point ; also dredged in 2-6 
fathoms. 



Cy there Packardi, G. S. Brady (PI. II. fig. 19). 

Cy there Packardi, Brady, " Ostracoda of Challenger Expedition," p. 88, pi. xix. fig. 2 a-d. 

The types of this species were found sparingly in a " Challenger" dredging from off 
Booby Island. 

There are two distinct forms of shell, probably belonging respectively to the two 
sexes. One of these is well figured in the "Challenger" Report. This I believe to be the 
female; the other is figured in this paper. It differs from the "Challenger" form in 
being less tumid, and in having a much more pronounced production of the shell at the 
postero-ventral angle. The surface is marked with very prominent longitudinal ribs, which 
in the female are indicated only faintly, or not at all. The male is smaller, and has the 
hinge-tubercle much more strongly marked than in the opposite sex. 

Habitat. — Noumea, in shore-sand from Porcheron's Beach and Artillery Point, also 
dredged in 2-6 fathoms ; Suva, mud-flats between tide-marks ; Levuka and Vuna Point, 
Taviuni, between tide-marks; Sava-Sava Bay, 4 fathoms; Rambe Island, between tide- 
marks ; Loma-Loma, in shore-sand; Lufi-Lufi and Apia, Upolu, in shore-pools. 

Cythere deltoides, n. sp. (PL II. figs. 17, 18). 

Shell, seen from the side, oblong, quadrangular, highest in front, height equal to 
rather more than half the length. Anterior extremity wide and moderately rounded, 
posterior much narrower, sharply produced below the middle into an angulated and 
irregularly dentated promontory, above which it slopes steeply backwards, the middle of 
the slope broken by a sharp angular projection; dorsal margin gibbous in front, thence 
sloping backwards in a sinuous curve, and terminating in a sharp angle ; ventral margin 
straight or more or less irregularly sinuated. Seen from above, the outline is irregular — 
doubly triangular — a large anterior triangle whose base forms the widest part, and a 
much smaller posterior triangle applied by its base to the larger ; anterior extremity 
rectangularly truncated ; the lateral margins diverging strongly and very irregularly to 
the base of the larger triangle, where they suddenly dip inwards at a right angle, running 
with a wide, much dentated sweep to the posterior extremity, which is narrower than the 
anterior and emarginate in the middle. Surface of the shell marked with numerous 
subcircular pittings, hinge-tubercle conspicuous, a more or less distinct ridge running 
obliquely across the valve from near the middle of the anterior margin to the posterior 
dorsal angle, the anterior portion of the valve ending precipitously in a transverse ridge 
at the posterior fourth. Length '66 mm. 

Habitat. — Port of Noumea, in shore-sand and near Artillery Point, and dredged in 
2-6 fathoms ; Lufi-Lufi, Upolu, reef and shore-pools ; Apia, Upolu, reef and shore-pools. 



502 DR G. S. BRADY ON 

Cythere prava, Baird. 

Cythere prava, Brady, "Ostracoda of Challenger Expedition," p. 92, pi. xxii. fig. 4 a-/. 

Habitat. — Suva, inside reef; Levuka, between tide-marks; Sava-Sava Bay, 4 fathoms; 
Vuna Point, Taviuni, between tide-marks; Mango Island, fringing reef; Rambe - Island, 
between tide-marks; Loma-Loma, in shore-sand ; Apia, Upolu, reef and shore-pools. 

This is the species noted in the " Challenger " Report as being found in dredgings 
from the Admiralty Islands. It appears to be a common form among the South Sea 
Islands, but differs a good deal from the European type. 

Cythere rectangularis, G. S. Brady. 

Cythere rectangularis, Brady, Les Fonds de la Mer, vol. i. p. 153, pi. xviii. figs. 13, 14 ; Linnean Soc. 
Journal (Zoology), vol. xix. p. 310, pi. xl. figs. 7-9. 

Habitat. — Noumea, Porcheron's Beach ; Levuka, between tide-marks ; Vuna Point, 
Taviuni, between tide-marks ; Rambe Island, between tide-marks ; Loma-Loma, in shore- 
sand. 

The specific name rectangularis is given in the "Challenger" Report as a synonym 
of Audei. This is a mistake, the shell figured as C. audei being quite distinct. The 
South Sea specimens here noticed seem to be certainly identical with those found in 
Ceylon, but they present a rather remarkable character, which is not visible in the Ceylon 
specimens, — the presence on the ventral margin, near the posterior extremity, of two 
very faint squamous dentations. This, however, though usual, is not visible on all 
shells. 

Cythere Goujoni, G. S. Brady. 

" Cythere Goujoni, Brady, " Ostracoda of Challenger Expedition," p. 96, pi. xxv. fig. 7 a-g. 
A widely distributed, and on that account, perhaps, a variable species. It has already 
been recorded from Ceylon, Hong Kong, China, Port Jackson, and Booby Island. 
Habitat. — Port of Noumea, 3-6 fathoms. 

Cythere infandibidata, n. sp. (PI. II. figs. 15, 16). 

Shell, seen from the side, oblong, subquadrangular, nearly equal in height throughout, 
height equal to half the length. Anterior extremity well rounded, divided into numerous 
short, blunt teeth; posterior produced and angulated a little below the middle, thence 
sloping with a very slight inward curve upwards and downwards ; ventral margin almost 
straight, slightly encroached upon in the middle by the central protuberance of the shell, 
rounded off in front, angulated behind ; dorsal margin parallel with the ventral, sinuous, 
joining the posterior margin at an obtuse angle ; elevated over the anterior hinge. Seen 
from above, the shell has somewhat the shape of a funnel with a large triangular prominence 
in the middle of the wide end ; greatest width near the hinder extremity, and equal to 
about two-thirds of the length ; from this point the lateral margins approximate with a 
convex curve until near the front, which is formed by a wide obtusely- rounded, sub- 
truncate process ; backvvardly the sides converge almost at a right angle towards the 



OSTRACODA COLLECTED TN THE SOUTH SEA ISLANDS. 503 

median line, then sloping sharply backwards and forming a wide acutely-pointed process. 
Shell-surface covered with irregularly shaped, rather large pits, central portion very 
tumid towards the ventral margin. Length '77 mm. 

Two or three specimens only were found between tide-marks at Vuna Point, Taviuni, 
Fiji. 

Cythere labiata, n. sp. (PI. II. figs. 20, 21). 

Shell, seen from the side, oblong, subquadrangular, highest in front, height equal to 
more than half the length. Anterior extremity well rounded, crenated ; posterior only 
moderately rounded, obscurely angulated above, rounded off below, bordered with 
strong teeth, which are longest below the middle ; dorsal margin forming an angular eleva- 
tion in front, thence sloping gently in a broken, tuberculated line backwards ; ventral 
straight, rounded off at each extremity, divided into closely-packed, rounded teeth. 
Seen from above, oblong, twice as long as broad, scarcely at all tapered at the extremities, 
which are wide and truncated ; edges everywhere much broken and toothed. The 
surface of the valves is tuberculated indistinctly for the most part, but on the posterior 
half shows an imperfectly linear arrangement of some of the tubercles ; the central 
area is moderately convex, and is separated from the surrounding dentated margin by a 
shallow furrow, extending round the shell, except on the dorsal margin ; the hinge-tubercle 
very large. Length 7 mm. 

Habitat. — Levuka, Fiji, between tide-marks. One specimen only. 

Cythere ichthyoderma, n. sp. (PL II. figs. 22, 23). 

Shell oblong, subquadrangular, highest over the anterior hinge, height equal to half 
the length. Anterior extremity well rounded, posterior subtruncate, only slightly 
rounded, much narrower than the anterior ; dorsal margin forming a hump over the 
anterior hinge, thence sloping gently and almost in a straight line backwards ; ventral 
straight. Seen from above, compressed, subovate, nearly thrice as long as broad, with 
very wide, equal, truncated extremities ; for the greater part the sides are nearly parallel, 
but at a short distance from each extremity they slightly converge. The surface of the 
valves is smooth, forming a convex area, encircled on all sides, but more especially in 
front and behind, by a wide, thickened lip. In well-marked specimens the central area 
shows, running obliquely across it in a longitudinal direction, two narrow squamous or 
tuberculated ridges, but one or both of these may be very faintly marked or even altogether 
wanting. The margins are dentated or spinulose at almost all points, except at the anterior 
part of the dorsum, just behind the hinge-tubercle ; on the anterior and posterior margins 
the teeth are generally short and blunt, and point .directly forwards and backwards ; while 
on the dorsal and ventral margins they are squamous in character, and have their points 
directed backwards. Length '77 mm. 

Habitat. — Port of Noumea, 3-6 fathoms; Suva, mud-flats and inside reef; Sava-Sava 
Bay, Vanua Levu, 4 fathoms ; Vuna Point, Taviuni, between tide-marks ; Rambe Island, 
between tide-marks ; Lufi-Lufi, Upolu, reef and shore-pools. 



504 DR G. S. BRADY ON 

The figures here given show a shell with the spinous armature pretty well developed, 
but in some the squamous, pointed character of the central and dorsal spines is much 
more apparent. 

Cythere quadriserialis, n. sp. (PI. II. figs. 27, 28). 

Shell, seen from the side, oblong, subquadrate, much higher in front than behind, 
height equal to more than half the length. Anterior extremity broad and boldly rounded, 
posterior much narrower, subtruncate or only moderately rounded, dorsal margin sloping 
steeply backwards, ventral almost straight. Seen from above, the outline forms a very 
irregular oblong, more than twice as long as broad, and widest behind the middle; very 
slightly narrowed towards the extremities, which are wide and truncated, the anterior 
deeply emarginate, margins extremely spinous and irregular. The surface of the valves 
shows a central convex area, encircled, except on the dorsum, by a wide, thickened flange 
or lip, which is everywhere bordered by short, closely-set, blunt teeth ; the anterior margin 
terminates above in a strongly-developed spine or group of spines, and behind this, on 
the dorsal margin, are three widely-detached groups of very large and strong spines, each 
group consisting of about three coalescent spines ; the central area of the valve is marked 
by two oblique ridges, composed of semi-detached bosses or tubercles, the upper rib 
being divided into two portions by a median gap. Length "85 mm. 

Habitat. — Noumea, in shore-sand, and dredged in 3-6 fathoms. 

A very distinct and remarkable species, no two specimens of which are exactly alike. 
The specimen figured exhibits a strongly-developed dorsal armature, but the ridges of the 
central area are not so continuous or so well marked as in many. 

Cythere militaris, G. S. Brady (PI. II. figs. 24-26). 

Cythereis militaris, G. S. Brady, On new or imperfectly known species of Marine Ostracoda 
{Trans. Zool. Soc, vol. v. p. 385, pi. lxi. fig. 9, a-d). 

Shell of the female, seen from the side, oblong, subquadrangular, highest in front, 
height equal to more than half the length. Anterior extremity wide and well rounded, 
posterior also rounded, but narrower ; dorsal margin sloping in a right line from the front, 
ventral straight. The central portion of the valves is smooth, convex, and bears three 
longitudinal rows of strong, blunt spines ; the middle row extends almost the whole 
length of the shell, but is interrupted in the middle ; the upper and lower rows are much 
shorter, each about one-third of the length of the valve, and placed just within the mid- 
region of the dorsal and ventral margins. This central area is bordered in front and 
behind by a thick, encircling lip, and the entire circumference is fringed with strong 
spines, which on the anterior and ventral margins are usually short and squared, but on the 
posterior and dorsal margins, especially at the infero-posteal angle, are developed in fully 
grown shells into long, sharp, curved spines. Seen from above, the outline is oblong, 
widest behind the middle, with broad, truncated, and spinous extremities. The shell of 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 505 

the male (?) is not so high as that of the female, and the posterior extremity is almost 
rectangularly truncate. Length (of both sexes) 1 mm. 

Habitat. — Suva Bay, 12 fathoms. My cabinet contains a good series of this fine 
species, from a dredging in Princess Charlotte Harbour, West Australia. An examina- 
tion of this series shows that the specimen figured in the Transactions of the Zoological 
Society (loc. cit.) is a young shell of the same species. I have, therefore, given here 
figures of the adult female form from the West Australian series. 

Genus Limnicy 'there, G. S. Brady. 

Limnicythere Fijiensis, n. sp. (PI. II. figs. 33, 34). 

Shell, seen from the side, reniform, slightly higher in front than behind, height equal 
to rather more than half the length. Anterior extremity wide and boldly rounded, 
posterior narrower, somewhat oblique (the obliquity looking downwards), and not very 
fully rounded ; both extremities more or less crenated, the crenations sometimes extending 
even on to the dorsum ; dorsal margin very slightly arched, ventral deeply incurved in 
the middle. Seen from above, the outline is irregularly wedge-shaped, more than twice as 
long as broad, the greatest width near the hinder end, anterior extremity blunt and 
emarginate in the middle, the sides gradually diverging to near the posterior extremity, 
then converging rather suddenly and in an irregular line to the wide and blunt extremity. 
Surface of the shell thickly covered with small oblong, impressed markings, and bearing a 
more or less distinct central tubercle with a surrounding depression ; within the ventral 
border is a not very distinct curved ridge, and in some specimens there are irregular 
faint undulated ridges on the central portion of the valve. Length 5 mm. 

Habitat. — Levuka, between tide-marks ; Rambe Island and Vuna Point, Taviuni, 
between tide-marks ; Mango Island, pools on the fringing reef ; Loma-Loma, in shore- 
sand. 

I have had no opportunity of examining the soft parts of this species, all the 
specimens being merely empty shells ; but from the general characters of the shell, I 
entertain little doubt that it belongs to the genus Limnicythere. 

Genus Cytheridea, Bosquet. 
Cytheridea spinulosa, G. S. Brady. 

Cytheridea spinulosa, Brady, " Ostracoda of Challenger Expedition," p. 112, pi. xxxiii. fig. 6 a-d. 

This species, originally described from specimens taken at Mauritius, seems to be 
widely distributed in the Southern Hemisphere. Amongst the " Challenger " dredgings, 
it was found at Amboyna and in a deep dredging from the South Pacific. Amongst the 
gatherings here described, it occurs as follows : — At Noumea, in shore-sand, from 
Porcheron's Beach and Artillery Point, and in dredgings from 2 to 6 fathoms ; mud-flats 
at Suva ; at Sava-Sava Bay, Vanua Levu, 4 fathoms ; and at Eambe Island, between 
tide-marks. 

VOL. XXXV. PART II. (NO. 14). 4 



500 DR G. S. BRADY ON 

Cytheridea jiavescens, n. sp. (PI. II. figs. 29-32). 

Shell of the female, seen from the side, subreniform, greatest height in the middle 
and equal to half the length, wide and evenly rounded in front, obliquely rounded behind, 
the obliquity looking upwards ; dorsal margin gently arched, ventral slightly sinuated 
in front of the middle. Seen from above, ovate, well rounded behind and subacutely 
pointed in front, more than twice as long as broad. Shell-surface smooth, yellowish, with 
darker clouded patches, bearing numerous distant, circular papillae, and marked on the 
anterior, posterior, and ventral margins with short radiating hair-like lines. The shell of 
the male is more elongated, lower, and has the postero-ventral angle more pronounced- 
Length of the male, *8 mm. ; of the female, '75 mm. 

Habitat. — Port of Noumea, 2-6 fathoms ; Levuka, between tide-marks ; Sava-Sava 
Bay, Vanua Levu, 4 fathoms. 

Cytheridea consobrina, n. sp. (PI. III. figs. 5, 6). 

Shell of the male, seen from the side, oblong, subovate, rather higher in front than 
behind, height equal to half the length. Extremities rounded, the posterior somewhat 
flattened, and bearing at its lower end a single strong, backward-pointing spine ; dorsal 
margin slightly arcuate, sloping gently downwards from near the front ; ventral margin 
almost straight. Seen from above, the outline is ovate, about twice and a half as long as 
broad, and with subparallel sides ; obtusely pointed in front, rounded off behind. Shell- 
surface marked with closely-set subcircular excavations. Length 1 mm. The shells of 
the two sexes are alike in size, but that of the female is higher in proportion, and its 
posterior half is very tumid. 

This species is in shape and general appearance of both sexes not unlike the common 
Cytheridea torosa of Europe, but differs very decidedly in the character of its surface 
markings, the fossae being much larger; the shell also is more elongated, and in the 
female much more tumid behind. It was found plentifully in shore-sand from near low- 
water mark at Noumea. 

Genus Loxoconcha, G. 0. Sars. 

Loxoconcha gracilis, n. sp. (PI. IV. figs. 24-36). 

Shell of the male, seen from the side, oblong-ovate, height equal to rather more than 
half the length. Anterior extremity well rounded, posterior obliquely truncated above 
the middle, rounded off below ; dorsal margin straight, ventral sinuated in front. Seen 
from above, compressed, ovate, more than twice as long as broad, widest in the middle, 
tapered gradually to the acute anterior extremity, abruptly towards the posterior, which 
is strongly mucronate. The shell of the female is shorter, and has a strongly-arched 
dorsum. Surface marked with closely-set rounded pits and a few distant circular papillae. 
Length of the male, '65 mm.; of the female, '55 mm. 

Habitat. — Noumea, in shore-sand, and dredged in 2-6 fathoms ; Suva, mud-flats and 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 507 

pools inside reef; Levuka, between tide-marks ; Sava-Sava Bay, Vanua Levu, 4 fathoms ; 
Vuna Point, Taviuni, between tide-marks ; Mango Island, fringing reefs ; Rambe Island, 
between tide-marks ; Loma-Loma, in shore-sand ; Lufi-Lufi ; Upolu, reef and shore- 
pools. 

This is perhaps the most abundant of the species met with in these gatherings, 
occurring more or less plentifully in almost all. 

Loxoconcha avellana, G. S. Brady. 

Loxoconcha avellana, Brady, " Ostracoda of Challenger Expedition," p. 117, pi. xxviii. fig. 1 a-f. 
Habitat. — Port of Noumea, 3-6 fathoms. 

Loxoconcha honoluliensis, G. S. Brady. 

Loxoconcha honoluliensis, Brady, "Ostracoda of Challenger Expedition," p. 117, pi. xxviii. fig. 6 
a-f 
Habitat. — Noumea, in shore-sand, and dredged in 2-6 fathoms ; Suva, mud-flats and 
pools inside reef ; Apia and Lufi-Lufi, Upolu, reef and shore-pools. 

Loxoconcha australis, G. S. Brady. 

Loxoconcha australis, Brady, " Ostracoda of Challenger Expedition," p. 119, pi. xxviii. fig. 5 a-f 
and pi. xxix. fig. 3 a-d. 
Habitat. — Noumea, in shore-sand, and dredged in 2-6 fathoms. 

Loxoconcha fumicom, G. S. Brady. 

Loxoconcha pumicosa, Brady, " Ostracoda of Challenger Expedition," p. 118, pi. xxviii. fig. 2 a-d. 

Habitat. — Noumea, in shore-sand, and dredged in 2-6 fathoms; Suva, inside reef; 
Vuna Point, Taviuni, between tide-marks ; Apia and Lufi-Lufi, Upolu, reef and shore- 
pools. 

Loxoconcha alata, G. S. Brady. 

Loxoconcha alata, Brady, Ann. and Mag. Nat. Hist., ser. 4, vol. ii. (1868) p. 223, pi. xiv. figs. 
8-1 3 (not Loxoconcha alata of the " Challenger " Report, which is Loxoconcha gibbera, Brady). 

Habitat. — Noumea, in shore-sand near Artillery Point ; Sava-Sava Bay, Vanua Levu, 
4 fathoms. 

Loxoconcha anomala, G. S. Brady. 

Loxoconcha anomala, Brady, "Ostracoda of Challenger Expedition," p. 123, pi. xxvii. fig. 5 a-d. 
Habitat. — Port of Noumea, 3-6 fathoms ; Levuka, between tide-marks. 

Loxoconcha dorso-tuberculata, G. S. Brady. 

Normania dorso-tuberculata, Brady, Trans. Zool. Soc, vol. v. p. 383, pi. lxi. figs. 14 a-g. 
Habitat. — Suva, mud-flats ; Levuka, between tide-marks ; Vuna Point, Taviuni, 
between tide-marks ; Mango Island, fringing reef ; Loma-Loma, in shore-sand. 



508 DR G. S. BRADY ON 

Loxoconcha gibbera, G. S. Brady (PI. IV. figs. 27, 28). 

Loxoconcha gibbera, Brady, Linn. Soc. Journal, "Zoology," vol. xix. p. 312, pi. xl. figs. 19-21. 

Habitat. — Mango Island, fringing reef. 

The figures given with the original description of this species are so inaccurate that I 
have drawn it afresh for this memoir. The illustrations now given may be taken as 
good representations of the type. 

Genus Xestoleberis, G. 0. Sars. 
Xestoleberis curta, G. S. Brady. 

Xestoleberis curta, Brady, " Ostracoda of Challenger Expedition," p. 126, pi. xxxi. fig. 6 a-d. 
Habitat. — Noumea, in shell-sand, and dredged in 2-6 fathoms ; Mango Island, 
fringing reef ; Lufi-Lufi, Upolu, reef and shore-pools. 

Xestoleberis variegata, G. S. Brady. 

Xestoleberis variegata, Brady, "Ostracoda of Challenger Expedition," p. 129, pi. xxxi. fig. 8 a-g. 

Habitat. — Noumea, dredged in 2-6 fathoms ; Suva, mud-flats ; Levuka, between tide- 
marks ; Sava-Sava Bay, Vanua Levu, 4 fathoms ; Yuna Point, Taviuni, between tide- 
marks ; Eambe Island, between tide-marks ; Loma-Loma, in shore-sand ; Apia and Lufi- 
Lufi, Upolu, reef and shore-pools, and in shore-sand. 

Shells which I refer to this species are very abundant in almost all these gatherings. 

Xestoleberis granulosa, G. S. Brady. 

Xestoleberis granulosa, Brady, " Ostracoda of Challenger Expedition," p. 125, pL xxx. fig. 5 a-d. 
Habitat. — Port of Noumea, 3-6 fathoms. 

Xestoleberis tumefacta* G. S. Brady (PI. III. figs. 7, 8). 

Xestoleberis tumefacta, Brady, "Ostracoda of Challenger Expedition," p. 128, pi. xxxi. fig. 4 a-d. 
Habitat. — Noumea, 2-4 fathoms. 

Xestoleberis gracilis, n. sp. (PI. III. figs. 9, 10). 

Shell, seen from the side, elongated, greatest height in the middle and equal to less 
than half the length. Anterior extremity much depressed, narrow and almost angular, 
posterior rounded; dorsal margin sloping with a gentle curve in front, evenly arched 
behind, ventral margin almost straight. Seen from above, the shell is ovate, widest 
behind the middle, width equal to fully half the length, tapered from the middle and 
acutely pointed in front, widely rounded behind. Shell quite smooth. Length "42 mm. 

Habitat. — Lufi-Lufi, Upolu, reef and shore-pools. 

Genus Cytherura, G. 0. Sars. 

Cytherura marcida, n. sp. (PI. III. figs. 24, 25). 

Shell of the male (?), seen from the side, rhomboidal, height the same throughout, and 

* The figures here given are taken from a specimen of extreme tumidity, but which seems to possess no characters 
sufficient to separate it from the ordinary form. 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 509 

equal to half the length. Anterior extremity rounded and somewhat oblique, posterior 
having a broad median beak, above which it slopes steeply and with a slight convexity to 
the dorsum ; below it is obliquely truncated, the obliquity looking downwards, and 
rounded off at the ventral angle ; dorsal margin straight in front, but elevated behind, 
and forming an angular hump ; ventral sinuated in the middle. Seen from above 
irregularly lozenge-shaped, compressed, widest near the middle, more than twice as long 
as broad; anterior extremity broad, only slightly rounded; posterior very wide, truncated, 
and having a blunt median emarginate process ; sides converging gradually and with a 
somewhat sinuous outline to the front, and deeply excavated towards the posterior 
extremity. Shell-surface very rugged, with irregular but not very prominent flexuous 
longitudinal ribs, which are connected transversely at intervals so as to form a rough reti- 
culation ; the anterior hinge-tubercle is large, polished, and conspicuous. There are two 
distinct forms of this shell, probably sexual. The second form differs from that described 
above, in having a much more regular surface with less rugged sculpture, a more even out- 
line, and no angular hump on the dorsum. This I take to be the female. Length "6 mm. 
Habitat. — Suva, reef ; Levuka, between tide-marks ; Loma-Loma, in shore-sand ; 
Apia, Upolu, reef and shore-pools. 

Cytherura entomon, n. sp. (PI. III. figs. 26, 27, 27a). 

Shell, seen from the side, subrhomboidal, height equal to rather more than half the 
length. Anterior extremity very oblique, slightly rounded, the obliquity looking upwards ; 
posterior broadly beaked above the middle, the apex of the beak emarginated, below this 
the extremity slopes sharply away backwards, forming a continuous line with the ventral 
margin; dorsal margin rugged, broken by numerous small blunt projections; ventral 
convex. Seen from above, the outline is jagged and very irregular, more or less ovate, 
obtusely pointed in front, more produced and acute behind, width equal to about one 
half the length ; behind the middle there is on each side a conspicuous club-like process 
directed transversely outwards. Shell-surface marked with indistinct oblong impressions, 
their long diameters directed transversely ; towards the posterior extremity there is in 
some specimens a transverse series of about 5 or 6 backwardly-directed spines. Length 
'5 mm. 

Habitat. — Port of Noumea, 3-6 fathoms ; Sava-Sava Bay, Vanua Levu, 4 fathoms. 

The description applies to the largest and most rugged specimens; but others occur 
in which the surface is not nearly so rough, the lateral outline more regular, the strong, 
transverse processes being absent, and the posterior extremity produced in a linear fashion. 
A shell of this kind is represented at fig. 27a. 

Cytherura scutellata, n. sp. (PI. III. figs. 30, 31). 

Shell, seen from the side, subrhomboidal, highest in the middle, height equal to more 
than half the length. Anterior extremity rounded, placed entirely below the middle line ; 



510 DR G. S. BRADY ON 

posterior produced in the middle into a large, wide, and truncated beak ; dorsal margin 
boldly arched, sloping steeply behind and more gradually in front; ventral almost straight, 
rounded off at each end. Seen from above, the outline is elongated, subhexagonal, with 
straight, parallel sides, which converge abruptly towards the extremities; anterior ex- 
tremity obtuse, subtruncate, posterior produced, tapering and pointed at the apex. 
Surface of the valves deeply excavated into a few (about twelve) polygonal, saucer-like 
cavities, each with an elevated nodule in its centre. Length *43 mm. 

Habitat. — Levuka, between tide-marks. I have this species also from Princess 
Charlotte Harbour in West Australia. 

This resembles so closely the well-known European species Cytherura cellulosa, 
(Norman), that I at first thought the two to be identical. But C. cellulosa is more 
angular in outline, has no distinct beak, and the surface excavations are much smaller 
and more numerous than in C. scutellata. 

Genus Cytheropteron, G. 0. Sars. 

Cytheropteron coccoides, n. sp. (PL III. figs. 20, 21). 

Shell, seen from the side, elongated, subtrapezoidal, nearly thrice as long as high ; 
height behind and in front equal. Extremities equal, depressed, produced and subangular 
below the middle, the anterior angle somewhat rounded ; dorsal margin almost straight 
in the middle, sloping with a steep curve at each end ; ventral slightly sinuated in front, 
convex behind. Seen from above, oval, not quite twice as long as broad, widest in the 
middle ; lateral margins evenly and boldly arcuate, extremities equal and almost rounded. 
Surface smooth, a slightly produced flange running round the valves, except on the dorsal 
margin. Length '46 mm. 

Habitat. — Mango Island, fringing reef. 

A very near ally of this species is a European one, C. humile, Brady and Norman, 
but C. coccoides is more depressed fore and aft, is more tumid, more pointed in front 
when seen dorsally, and has a distinctly papillose surface. 

Cytheropteron rude, n. sp. (PI. III. figs. 15-17). 

Shell, seen from the side, subrhomboidal, highest in front, height equal to more than 
half the length. Anterior extremity very wide, obliquely rounded, the obliquity looking 
upwards ; posterior tapered, produced in the middle to a subacute angle ; dorsal margin 
high in front, sloping with a bold but irregular curve to the posterior extremity, ventral 
margin irregularly sinuous, angulated behind. Seen from above, the outline is hexagonal, 
widest in the middle, the width equal to nearly two-thirds of the length, lateral margins 
for the middle third of their course straight and parallel, converging with a steep slope to 
the acute anterior extremity, dipping at a right angle behind the middle, then sloping 
sharply to the posterior extremity, which is acutely mucronate. End view very irregular, 
wide at the base, which is produced laterally into rounded prominences, and is thrice 
emarginate in the middle ; lateral margins very deeply sinuated, apex wide and emar- 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 511 

ginate in the middle. Surface rugged, irregularly undulated and pitted ; a longitudinal 
alaeform process just within the ventral margin, but not strongly developed. Length 
"43 mm. 

Habitat. — Sava-Sava Bay, Vanua Levu, 4 fathoms. 

Cytheropteron longicaudatum, n. sp. (PI. III. tigs. 18, 19). 

Shell, seen from the side, oblong, nearly equal in height throughout, height equal to 
about half the length. Anterior extremity obliquely rounded, the obliquity looking 
upwards ; posterior produced into an acutely- tapering median triangular beak of great 
length ; dorsal margin convex before and behind, with a very deep median sinus ; ventral 
convex and rather sinuous, terminating in an abrupt angle behind. Seen from above, the 
outline is hastate, widest near the middle, width equal to two-thirds of the length ; from 
the widest point the sides converge with a bold convexity forwards, forming a lancet-shaped 
front ; posterior extremity forming a wide, tapering and very acute triangular process. 
Surface of the valves deeply furrowed across the middle, bearing also a more or less 
distinct central tubercle in front of the groove, and several obscurely radiating ribs with 
small fossse in the interspaces. Length '57 mm. 

Habitat. — Suva, mud-flats ; Levuka, between tide-marks ; Sava-Sava Bay, Vanua 
Levu, 4 fathoms ; Rambe Island, between tide-marks ; Loma-Loma, in shore-sand. 

Cytheropteron guttatum, n. sp. (PL IV. figs. 29, 30). 

Shell, seen from the side, subovate, highest in front of the middle, height equal to 
about two-thirds of the length. Anterior extremity wide and rounded, posterior sub- 
truncated, slightly produced below the middle ; dorsal margin well arched, obscurely 
angular over the anterior hinge ; ventral straight, slightly sinuated in front and behind ; 
lateral ala moderately prominent, and rounded off at each extremity. Seen from above, 
the outline is lozenge-shaped, widest behind the middle, width equal to two-thirds of the 
length ; anterior extremity very obtuse, rounded ; sides sloping with a bold convexity 
backwards, and near the posterior extremity dipping suddenly in a hollow curve to the 
extremity, which is wide and truncate, but narrower than the anterior. Surface smooth, 
marked throughout with closely-set circular, impressed puncta ; hinge-tubercle polished 
and conspicuous. Length '5 mm. 

Habitat. — Noumea, dredged in 2-6 fathoms. 

Cythcropteron (?) trilobites, n. sp. (PI. III. figs. 22, 23). 

Shell, seen from the side, oblong, quadrangular, highest in front, height equal to 
about half the length. Anterior extremity wide, obliquely subtruncated, only moderately 
rounded ; posterior produced above the middle into a triangular beak, excavated below 
the middle ; dorsal margin sloping rather steeply from the front, and showing two deep 
angular sinuations, one in front of and the other behind the middle ; ventral nearly 



512 DR G. S. BRADY ON 

straight, ending abruptly behind in a rectangular process. Seen from above, the outline is 
much like that of a trilobite, very wide and rounded in front, and narrowing a little 
towards the posterior extremity, which is wide and truncated, with a median triangular 
beak ; width equal to more than two-thirds of the length. There is a slight median 
prominence on the anterior margin, and a more or less distinct constriction of the lateral 
margins in front of the middle, each side ending behind in an acutely-produced angle. 
The surface of the shell is very irregularly rugose and nodulated. End view quadri- 
lateral, base very wide and prominent in the middle, apex wide and obliquely truncated, 
sides moderately convex, width greater than the height. Length *5 mm. 
Habitat. — Banc de l'Aiguille, New Caledonia, dredged in 2-3 fathoms. 

Genus Cytherideis, Jones. 
Cytherideis baculoides, n. sp. (PI. III. figs. 11, 12). 

Shell, seen from the side, elongated, oblong, equal in height throughout, height equal 
to not much more than one-fourth of the length. Anterior extremity suddenly depressed, 
rounded off, almost angulated below ; posterior evenly rounded ; dorsal and ventral 
margins parallel and perfectly straight. Seen from above, compressed, fusiform, more 
than four times as long as broad, acuminate in front, rounded off behind. Shell pellucid, 
smooth, with a few minute, scattered, dot-like hairs. Length '75 mm. 

Habitat. — Levuka, between tide-marks ; Sava-Sava Bay, Vanua Levu, 4 fathoms. 

Fam. Paradoxostomatid^:. 

Genus Paradoxostoma, Fischer. 
Paradoxostoma overturn, n. sp. (PI. III. figs. 32, 33). 

Shell, seen from the side, ovate, highest behind the middle, height equal to half the 
length. Anterior extremity narrow, evenly rounded ; posterior obscurely angulated in 
the middle ; the dorsal margin is boldly arched, and forms a continuous curve to the 
angulation of the posterior extremity, sloping steeply behind and very gently in front ; 
ventral margin very slightly sinuated in front, boldly convex behind, and continuous with 
the posterior extremity. Seen from above, compressed, fusiform, widest in the middle, 
more than three times as long as broad, extremities equal and acuminate. Surface 
smooth, colour greenish, with a dark mottled band in the middle. Length '5 mm. 

Heibitat. — Between tide-marks, Vuna Point, Taviuni ; and Levuka, north of the town. 

Paradoxostoma NoveB-Caledonice, n. sp. (PL IV. fig. 19). 

Shell, seen from the side, elongated, subovate, highest rather behind the middle, 
height equal to rather less than half the length. Anterior extremity narrow, evenly 
rounded ; posterior scarcely rounded, subangular ; dorsal margin rather boldly arched, 
highest behind the middle, thence sloping with a steep curve backwards and much more 



OSTKACODA COLLECTED IN THE SOUTH SEA ISLANDS. 513 

gradually to the front; ventral margin sinuated in the middle, up-curved at each extremity. 
Seen from above, compressed, ovate, widest in the middle, sides arcuate and tapering 
evenly to the extremities, which are pointed and nearly equal ; not quite four times as 
long as broad. Shell smooth and pellucid, marked with opaque patches. Length *55 mm. 
Habitat. — Port of Noumea, 3-4 fathoms. 

Paradoxostoma retusum, n. sp. (PI. IV. fig. 20). 

Shell, seen from the side, oblong, flexuous, highest behind the middle, height equal to 
nearly half the length. Anterior extremity narrow, evenly rounded ; posterior produced 
above the middle into a rounded beak, from which it slopes downwards and forwards with 
a full curve ; dorsal margin boldly arched, sinuated in front of the posterior extremity ; 
ventral rather deeply sinuated in the middle, behind which it forms a compressed and 
very convex marginal flange. Seen from above, compressed, widest in the middle, quite 
four times as long as broad; lateral margins evenly curved, anterior extremity obtuse, 
posterior acute. Shell pellucid, smooth, marked with opaque patches. Length '52 mm. 

Habitat. — Apia, Upolu, pools on inner reef. 

Section II. MYODOCOPA. 

Fam. CYPRIDINIDiE. 

Genus Philomedes, Lilljeborg. 

Philomedes vellicata, n. sp. (PI. IV. figs. 9, 10). 

Shell, seen from the side, oblong, subovate, greatest height in the middle, and equal to 
half the length. Anterior extremity narrow, forming a rounded beak, beneath which is a 
rather shallow notch ; posterior obliquely truncate, the obliquity looking upwards ; dorsal 
margin moderately and evenly arched, rounded off in front, obscurely angular behind ; 
ventral evenly convex, rounded at both ends. Seen from above, the outline is ovate, much 
compressed, about three times as long as broad, with subparallel, slightly arcuate sides, 
which are slightly constricted behind the middle ; obtuse in front, truncate behind, with a 
stout median prominence. Surface of the valves smooth, very finely punctated with linear 
dots, and marked behind the middle with a curved transverse furrow, which extends from 
the dorsal margin to below the middle of the valve. Length 1*1 mm. 

Habitat. — Suva, pools on reef ; Levuka, between tide-marks. 

Genus Pleoschisraa* no v. gen. 

Shell very dense, surface pitted, smooth or tuberculated; seen from the side subcircular, 
with a slight depression in place of a notch. 

Pleoschisma robusta, n. sp. (PI. IV. figs. 13, 14). 

Shell, seen from the side, nearly circular, highest in the middle, height equal to more 

* srAs'os", full ; <sxw(x.yi, a cleft. 
VOL. XXXV, PART II. (NO. 14). 4 P 



51-t DR G. S. BKADY ON 

than three-fourths of the length ; anterior extremity rounded and slightly sinuated 
(scarcely notched) below the middle ; posterior margin narrow and rectangularly truncate ; 
dorsal margin boldly arched, its hinder half scarcely at all curved ; ventral very strongly 
arcuate. Seen from above, ovate, with wide obtuse extremities ; greatest width in the 
middle, and equal to less than half the length. Surface marked with closely-set circular 
punctations, and in old shells covered with a dense reddish-brown incrustation. Length 
'9 mm. 

A shell presenting much the same general characters as the above, but smaller and 
more angular in outline and with a more distinct notch, occurred in the same gathering. 
This may perhaps be the male of P. robusta. I have had no opportunity of examining 
in detail the soft parts of the animal, all my specimens being dried shells and containing 
little or no remains of the internal structures ; but the fragments which I have seen agree 
in general character with Cypridina. 

Habitat. — Vuna Point, Taviuni, low-tide pools. 

Pleoschisma moroides, n. sp. (Plate I. figs. 23, 24). 

Shell, seen from the side, subcircular, height equal to three-fourths of the length. 
Anterior extremity wide, feebly rounded, almost flat, notch obsolete ; posterior narrower, 
rounded, slightly sinuated above and below ; dorsal and ventral margins moderately 
convex. Seen from above, broadly ovate, nearly equal in width throughout; extremities 
broad and rounded, the anterior rather the narrower of the two ; lateral margins 
moderately arcuate, width equal to four-sevenths of the length. Surface of the shell 
minutely punctated, and in old specimens raised into circular bosses; colour dark brown. 
Length 1*2 mm. 

Habitat. — Port of Noumea, dredged in 3-6 fathoms; Suva, inside reef; Levuka, 
between tide-marks ; Vuna Point, Taviuni, between tide-marks ; Mango Island, fringing 
reef. 

This curious species occurs not uncommonly in several of the localities above 
mentioned, but I have been unable to obtain more than very fragmentary specimens of 
the soft parts, which, so far as can be ascertained, closely resemble those of Cypridina. 

Pleoschisma reticulata, n. sp. (PL IV. figs. 11, 12). 

Shell, seen from the side, subcircular, greatest height situated in the middle and 
equal to three-fourths of the length. Anterior extremity broadly rounded, slightly 
sinuated below the beak, which is short and obtuse; posterior margin narrower, rounded, 
obs'eurely angular at its junction with the dorsum, but rounded off below ; dorsal and 
ventral margins boldly convex. Seen from above, broadly ovate, widest in the middle, 
width equal to three-fifths of the length ; extremities obtuse, rounded, rather more 
tapered in front than behind ; shell-surface smooth, marked with a delicate reticulated 
pattern. Length '57 mm. 



OSTEACODA COLLECTED IN THE SOUTH SEA ISLANDS. 515 

Only one specimen seen, and the locality in which it occurred was, unfortunately, not 
noted. 

Genus Asterope, Philippi. 
Asterope cylindrica, n sp. (PI. IV. figs. 7, 8). 

Shell, seen from the side, oblong, oval, of equal height throughout, height equal to 
two-fifths of the length. Extremities equal and well rounded, the anterior only slightly 
sinuated below the beak ; dorsal and ventral margins quite straight. Seen from above, 
the outline is elongated, ovate, three times as long as broad; extremities nearly equal, 
very obtusely pointed. Surface of the valves quite smooth. Length 1*3 mm. 

Habitat. — Suva, inside reef. 

Asterope australis, n. sp. (PI. IV. figs. 1, 2). 

Shell, seen from the side, broadly ovate, rather higher behind than in front, height 
equal to about three-fifths of the length ; extremities rounded ; notch of moderate 
depth, beak subacute ; dorsal and ventral margins moderately and equally convex. Seen 
from above, ovate, pointed in front, moderately rounded behind, widest in the middle ; 
width equal to two-fifths of the length. Surface smooth. Length 2*1 mm. 

Habitat. — Noumea, dredged, 2-4 fathoms; Suva, inside reef; Mango Island, fringing 
reef ; Apia, Upolu, reef and shore-pools. 

Genus Streptoleberis, nov. gen. 

Shell, seen from the side, elongated, flexuous ; beak much produced forwards, the 
notch being on the ventral aspect of the shell; posterior extremity narrower, produced 
into a pointed terminal beak. Animal unknown. 

Streptoleberis crenulata, n. sp. (PL IV. figs. 3, 4). 

Shell, seen from the side, irregularly lozenge-shaped, height equal to half the length, 
greatest in the middle. Anterior extremity produced in the middle line into a rounded 
beak, which is crenated at the apex ; posterior narrowed, also produced in the median line 
into a sharp triangular beak ; dorsal margin very slightly arcuate in the middle, sloping 
with a steep curve to the front, angulated behind, and thence sloping very abruptly to 
the terminal beak ; ventral margin slightly convex, having a wide but shallow and 
angular notch in front, up-curved behind to the posterior extremity. Seen from above, 
ovate, widest near the front, width equal to about two-fifths of the length, tapering 
rather abruptly to the front, which is sharply pointed ; posterior extremity rather 
broadly rounded; sides sinuous. Surface of the shell covered with small, circular 
impressions, and marked in a somewhat reticulated fashion with irregularly flexuous 
elevated ridges. Length 1*05 mm. 

Habitat. — Noumea, dredged in 2-4 fathoms. 



516 DE G. S. BRADY ON 

Only one example of this species was found in the Noumea dredging, but the genus 
was already familiar to me from specimens dredged in the North Atlantic, but not yet 
described. The very much produced extremities and the twisted form of the shell are 
quite characteristic. 

Genus Sarsietta, Norman. 
(British Association Report, 1868, p. 292.) 

Sarsiella scutyta, n. sp. (PI. I. figs. 17-20). 

Shell, seen from the side, subcircular, height and length nearly equal. Anterior 
extremity flattened, truncate, having a wide triangular prominence above, and a 
similar but less pronounced process below ; posterior extremity rounded and bordered 
with a more or less regular series of small nodular prominences; dorsal margin arched, 
sinuated at its junction with the posterior border; ventral convex, generally somewhat 
crenulated. Seen from above, the outline is subcuneiform, wide and truncated behind, 
with a prominent median beak, obtusely pointed in front, the sides parallel behind the 
middle, but converging gradually towards the front. Surface of the valves undulated, 
marked with closely-set small excavations, and having two stout flexuous ribs running in 
a longitudinal direction from near the triangular prominences of the anterior margin. 
These ribs are in some cases lost near the centre of the valve, and sometimes stretch over 
nearly its whole length, and there are often numerous smaller ridges running in a radial 
direction from the circumference of the shell inwards. Length 1 *4 mm. 

Habitat. — Noumea, dredged in 2-4 fathoms ; Levuka, between tide-marks ; Vuna 
Point, Taviuni, between tide-marks. 

This appears to be a not uncommon species of a group which, judging from the 
evidence of these gatherings, is much more strongly represented in the Southern than in 
the Northern Hemisphere. Almost nothing was seen of the soft parts of the animal. 
The very variable sculpturing of the shell— no two specimens being exactly alike 
in this respect — seems to depend partly on age and partly, perhaps, on sex. The 
figure 18, having been drawn from a gaping shell, gives an incorrect idea of its width. 
The description above given applies to specimens of the type figured in PI. I. figs. 17, 
18 ; but in a dredging from off Cap Bon Louis, New Caledonia, there occurred two 
specimens, differing very considerably from the types, but which from their general aspect 
and the close similarity of sculpture, appear to be, if not the same species, at any rate so 
closely related that I cannot find any satisfactory distinctive characters. One of these 
specimens is figured in PL I. figs. 19, 20. 

Sarsiella simplex, n. sp. (PI. IV. figs. 15, 16). 

Shell, seen from the side, almost circular, with a large median beak-like process 
behind ; length and height (exclusive of the beak) equal ; beak subtriangular, truncated at 
the apex. Seen from above, the outline is lozenge-shaped, widest in the middle, twice as 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 517 

long as broad; anterior extremity obtusely rounded; posterior tapered and subacute; 
lateral margin very boldly convex. Surface of the shell perfectly smooth. Length 1*05 
mm. 

Habitat. — Port of Noumea, 2-6 fathoms. 

Sarsiella rudis, n. sp. (PI. IV. figs. 5, 6). 

Shell, seen from the side, subcircular, with a prominent median beak ; height and 
length (exclusive of the beak) about equal. Anterior margin rounded, rather flat above 
the middle, posterior flattened and sloping steeply above the beak, sinuated below ; dorsal 
margin flattened, almost straight ; ventral boldly convex. Seen from above, the outline 
is hexagonal, with parallel, straight sides converging abruptly and equally to the ex- 
tremities, which are rather wide and obtuse ; width equal to two-thirds of the length. 
Surface of the shell devoid of regular sculpture, but vaguely ridged and undulated. 
Length '84 mm. 

Habitat. — Eambe Island, between tide-marks ; Suva, shallow water inside reef. 

Sarsiella foveata, n. sp. (PI. I. figs. 21, 22). 

Shell, seen from the side, almost circular, with a prominent beak ; height equal to 
about six-sevenths of the length (exclusive of the beak). Anterior extremity rounded, 
slightly prominent about the middle, posterior also rounded ; beak large, equal to one- 
fourth of the height of the shell, truncated at the apex ; dorsal and ventral margins boldly 
convex, dorsal sloping with a gentle curve towards the hinder end, ventral curve much more 
abrupt and almost angulated behind. Seen from above, the outline is elongated, sub- 
hexagonal, widest behind the middle, width equal to half the length ; anterior extremity 
broad and rounded, emarginate in the middle ; posterior sharply pointed ; the sides con- 
verge gradually from their widest point towards the front ; backwards the convergence is 
much more abrupt and sinuous ; the whole circumference is much jagged. End view 
subquadrangular, constricted in the middle. Shell-surface marked throughout with large 
and sharply-cut, deep, angular excavations; valves protuberant behind the middle, and 
forming towards the ventral margin an angular prominence. Length 1 *3 mm. 

Habitat. — Banc de 1' Aiguille, Noumea, 2-3 fathoms. One specimen only seen. 

Section III. PLATYCOPA. 

Fam. Cytherellid^e. 

Genus Cytherella, Jones. 

Cytherella semitalis, G. S. Brady. 

Cytherella semitalis, Brady, "Ostracoda of Challenger Expedition," p. 175, pi. xliv. fig. 2 a-e. 
Habitat. — Port of Noumea, 3-6 fathoms ; Suva-Suva Bay, Vanua Levu, 4 fathoms. 



518 DR G. S. BRADY ON 

Cijtherella (X) tumida, n. sp. (PL IV. figs. 21-23). 

Shell, seen from the side, somewhat oblique, subelliptical, rather lower in front than 
behind, height equal to more than half the length ; left valve much larger than the right, 
and overlapping greatly on the dorsal margin. Extremities rounded, the anterior some- 
what the wider ; dorsal and ventral margins only slightly convex, parallel. Seen from 
above, broadly ovate, widest behind the middle, width equal to about two-thirds of the 
length, obtusely pointed in front, broadly rounded behind, lateral margins boldly convex. 
Shell-surface quite smooth. Length "48 mm. 

Habitat. — One specimen only of this species was found in a gathering from reef-pools 
at Lufi-Lufi, Samoa. This specimen was destroyed in an unsuccessful attempt to find 
the contained animal. The generic reference must be considered only provisional, one 
important difference between this and the typical Cytherella being the larger size of the 
left valve ; whereas the valve of the right side is the larger in Cytherella. 

Cytherella cuneolus ? G. S. Brady. 

Cijtherella cuneolus ? Brady, Les Fonds de la Mer, vol. i. p. 192, pi. xix. figs. 18, 19. 

A shell, which probably belongs to this species, was found amongst shore-sand from 
Porcheron's Beach, Noumea. The specimen is, however, malformed, and the two valves 
differ considerably one from the other in shape and sculpture, so that I cannot assign it 
with certainty to this or any species. 

The following is a descriptive list of the gatherings in which the specimens have been 
found. The particulars in each case have been inserted from notes supplied to me by 
my brother, Dr H. B. Brady, F.R.S., to whom I am indebted for the material. In 
assigning localities to the different species, I have not thought it necessary in all cases to 
specify these localities with absolute accuracy, as, for instance, in the case of the several 
gatherings in or near the Port of Noumea, where the depth and physical conditions do not 
present any great variety. The multiplication of references to such localities could 
scarcely serve any useful end. 

New Caledonia. 

1. Noumea. — Porcheron's Beach, near the salt-flats ; brackish mud from pools about 
the mangrove trees, near or above high- water mark. 

2. Noumea. — Shore-sand near low water, head of bay, close to the road leading to 
Artillery Point ; reddish-brown muddy sand with stones, mollusc shells, fragments of 
Echini, Orbitolites, and Alveolinse. 

3. Port of Noumea. — 3-4 fathoms ; muddy sand, full of small Orbitolites and 
Alveolinse. 

4. Port of Noumea. — South side, off Artillery Barracks, 5-6 fathoms ; soft muddy 
sand, with mollusc-shells, whole and broken, and some coral. 

5. Near Noumea. — Between He Pore-Epic and shore ; weedy bottom, depth 2 fathoms. 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 519 

6. Near Noumea. — Off Cap Bon Louis, 4 fathoms ; weedy bottom. 

7. Near Noumea. — Banc de l'Aiguille, 2-3 fathoms ; weedy bottom, coral sand, with 
a few Orbitolites. 

Fiji. 

8. Suva. — Mud-flats between tide-marks ; fine muddy sand, with remains of Mollusca, 
Echini, and Polyzoa. 

9. Suva. — Inside reef, pools and shallows ; weedy bottom ; coral sand. 

10. Suva Bay. — 12 fathoms ; anchor-mud. 

11. Sava- Siva Bay, Vanua Levu. — 4 fathoms; anchor-mud. 

12. Levuka, Ovalau. — Beach north of the town ; sand from between tide-marks and 
from shore-pools ; coral sand, with Polytrema, Orbitolites, fragments of Mollusca, Echini, 
&c. 

13. Vuna Point, Taviuni. — Low- water pools and shore-sand; black volcanic sand, 
laden with organic fragments, Orbitolites, Polytrema, Diatomacese, &c. 

14. Mango Island. — From the boat-track on fringing reef, very shallow, about 1 
foot at low tide ; coral sand. 

15. Rambe Island. — Shore-sand and low- water pools; rough sand, with Orbitolites, 
Polytrema, &c. 

16. Loma-Loma, Vanua Mbalavu. — Sand from between tide-marks ; fine sand, with 
coral and sponge fragments and decaying vegetable matter. 

Samoa. 

1 7. Apia, Upolu. — Pools on inner barrier reef and shallows between reef and shore ; 
dead coral, with Mollusca, nullipores, &c. 

18. Lufj-Luji, Upolu. — Coral sand from reef and from pools and shallows — 3 or 4 
feet deep — between reef and shore. 

19. Luji-Lufi, Upolu. — Shore-sand from coast ; chiefly volcanic sand, with shell 
fragments, much worn. 

The following is a complete list of the species. The numerals refer to the places in 
which the species occurred, and correspond to those prefixed to the localities in the fore- 
going list : — 

LIST OF SPECIES. 
PODOCOPA. 

Cypridid^e. 

Phlyctenophora viridis, n. sp., 3, 11, 12, 13, 14, 15, 16, 17, 18, 19. 

rendformis, n. sp., 8, 14, 16, 18, 19. 
Pontocypris attenuata, G. S. Brady, 3, 5, 6, 17. 

„ gracilis, n. sp., 12, 15. 

„ sicula, n. sp., IL 

Anchistrocheles fumata, n. sp., 18. 



520 DR G. S. BRADY ON 

BAIRDIIDtE. 

Mucrocypris decora, G. S. Brady, 1, 2, 3, 4, 13. 
Bairdia simplex, G. S. Brady, 8, 13. 

„ tenera, G. S. Brady, 17, 18. 

„ amygdaloides, G. S. Brady, 3, 5, 6, 9, 12, 13, 14, 15, 16. 

„ Crosskeiana, G. S. Brady, 8, 17, 18. 

„ Woodwardiana, G S. Brady, 13. 

foveolata, G S. Brady, 1, 2, 3, 4, 5, 6, 7, 8, 11, 13, 14, 17, 18. 
Milne-Edwardsii, G. S. Brady, 5, 7, 9, 12, 13, 16, 17. 
Bairdia nodidifera, n. sp., 12. 

„ truncata, n. sp., 1, 17. 

„ ventricosa, n. sp., 2. 

„ tubercidata, G. S. Brady, 3. 

„ expansa, G. S. Brady, 5, 7, 8, 17, 18. 

„ hirsuta, G. S. Brady, 3. 

Cytheridje. 

Cythere demissa, G. S. Brady, 1, 3, 4, 5, 6, 7, 11, 12, 14, 15, 17. 
crenata, n. sp., 1, 2, 3, 4, 6, 7, 9, 11, 12, 14, 15, 17. 
,, ochracea, n. sp., 1, 2, 7. 
„ inflata, n. sp., 8, 11, 12, 13, 15, 16, 19. 
„ cuneolus, n. sp., 7, 14, 16. 
„ ovalis, G. S. Brady, 14. 
„ caudata, n. sp., 11. 
„ Scotti, n. sp., 5, 7. 
„ torticollis, n. sp., 2, 3, 5, 6, 7. 
„ Packardi, G. S. Brady, 1, 2, 3, 5, 6, 7, 8, 11, 12, 13, 15, 16, 17, 18. 

deltoides, n. sp., 2, 3, 4, 7, 17, 18, 19. 
„ infundibulata, n. sp., 13. 
„ prava, Baird, 9, 11, 12, 13, 14, 15, 16, 17. 
„ rectangularis, G S. Brady, 1,12, 13, 15, 16. 
„ Goujoni, G. S. Brady, 3, 4. 
„ labiata, n. sp., 12. 

„ ichthyoderma, n. sp., 4, 8, 9, 11, 13, 15, 18. 
„ militaris, G. S. Brady, 10. 
„ quadriserialis, n. sp., 2, 3, 4. 
Limnicythere Fijiensis, n. sp., 12, 13, 14, 15, 16. 
Cytheridea flavescens, n. sp., 3, 4, 11, 12. 
„ consobrina, n. sp., 2. 

„ spinulosa, G. S. Brady, 1, 2, 3, 4, 10, 11, 15. 

Loxoconcha gracilis, n. sp., 1, 2, 3, 8, 9, 11, 12, 13, 14, 15, 16, 18, 19. 
„ avellana, G. S. Brady, 3. 

Honoluliensis, G. S. Brady, 1, 2, 3, 4, 5, 6, 7, 8, 9, 17, 18. 
„ australis, G. S. Brady, 4, 6. 

pumicosa, G. S. Brady, 1, 2, 3, 5, 6, 7, 9, 13, 17, 18, 19. 
„ alata, G S. Brady, 11. 

„ anomala, G. S. Brady, 3, 12. 

„ dorso-tuberculata, G. S. Brady, 8, 12, 13, 14, 15, 16. 

„ (jibbera, G. S. Brady, 14. 

Xestoleberis cwrta, G. S. Brady, 1, 2, 5, 6, 7, 14, 18. 

variegata, G. S. Brady, 3, 4, 5, 6, 7, 8, 11, 12, 13, 15, 16, 17, 18, 19. 
„ gracilis, n. sp., 18. 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 521 

Xestoleberis granulosa, G. S. Brady, 3. 

„ tumefacta, G. S. Brady, 7. 

Cytherura longicaudata, d. sp., 11, 12, 13, 16. 
marcida, n. sp., 12, 16, 17, 18. 

„ entomon, n. sp., 3, 11. 

„ scutellata, n. sp., 12. 
Cytheropteron coccoides, n. sp., 14. 

„ rude, n. sp., 11. 

„ longicaudatum, 8, 11, 15, 16. 

„ guttatum, n. sp., 3, 4. 

„ trilobites, n. sp., 7. 

Cytherideis baculoides, n. sp., 11, 12. 

Paradoxostomatid^e. 

Paradoxostoma ovatum, n. sp., 12, 13. 

„ Novm Caledonice, n. sp., 3. 

„ retusum, n. sp., 17. 

MYODOCOPA. 

Cypridinim:. 

Philomedes vellicata, n. sp., 9, 12. 
Asterope australis, n. sp., 6, 7, 9, 13, 14, 17. 

„ cylindrica, n. sp., 9. 
Streptoleberis crenulata, n. gen. and sp., 6, 7. 
Pleoschisma robusta, n. gen. and sp., 13. 
„ reticulata, n. gen. and sp. (?). 

„ moroides, n. gen. and sp., 3, 6, 9, 12, 13 14. 

Sarsiella simplex, n. sp., 3, 6. 

„ foveata, n. sp., 7. 

„ rudis, n. sp., 9, 15. 

„ sculpta, n. sp., 5, 6, 12, 13. 

PLATYCOPA. 

CyTHERELLIDjE. 

Gytherella semitalis, G. S. Brady, 4, 10, 11. 
cuneolus ? G. S. Brady, 1 (?). 
„ tumida, n. sp., 18. 



EXPLANATION OF PLATES. 

Plate I. 

Phlyctenophora viridis. 
Fig. 1. Shell seen from left side 



above J 



x50. 

Pontocypris attenuata. 

3. Shell seen from left side I x 50 

4. ,, ,, above i 



VOL. XXXV. PART II. (NO. 14). 4 Q 



522 DR G. S. BEADY ON 

Pontocypris gracilis. 
Fig. 5. Shell seen from left side } 
„ 6. „ „ above / 

Pontocypris sicula. 
„ 7. Shell seen from left side 
„ 8. „ „ above 

Plilyctenoplwra reniformis. 
„ 9. Shell seen from left side 
„ 10. „ „ above 

Bairdia tenera. 

11. Shell seen from left side 

x 40. 



1 x40. 

?. 

1 x60. 



12. „ ,, above 



} 



Bairdia nodulifera. 

13. Shell seen from left side \ 

14. „ „ above I x 40. 

15. „ „ behind ) 

16. Posterior extremities of valves seen obliquely. 

Sarsiella sculpta. 

17. Shell seen from right side- 

18. „ „ above , 

19. „ „ right side 

20. „ „ below 

Sarsiella foveata. 

21. Shell seen from left side ) ,~ 

22. „ „ below J X 

Pleoschisma moroides. 

23. Shell seen from right side ) . „ 

24. „ above j 



Plate II. 

Bairdia truncata. 
Fig. 1. Shell seen from left side 
„ 2. ,, „ above 



}* 



65. 



Cythere inflata. 

3. Shell seen from left side 

4. „ „ above \ x 60. 
5- » » „ 

Cythere cuneolus. 

6. Shell seen from left side 

7. „ „ above 



Cythere ochracea. 
„ 8. Shell seen from left side 
„ 9. „ ,, above 

Cythere caudata. 
„ 10. Shell seen from left side 
„ 11. „ „ above 

Cythere ovalis. 
„ 12. Shell seen from left side x 50. 



} 

I x80. 



x80. 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 523 



I x80. 



50. 



Cythere nodulosa. 
Fig. 13. Shell seen from left side 
„ 14. „ „ above 

Cythere infundibulata 
,, 15. Shell seen from left side 
„ 16. ,, „ above 

Cythere deltoides. 
„ 17. Shell seen from left side 
„ 18. „ „ above 

Cythere Packard?'. 
„ 19. Shell seen from left side x 60. 

Cythere labiata. 
„ 20. Shell seen from left side 
„ 21. „ ,, above 



} 



x60. 



}« 



50. 



Cythere ichthyoderma. 
22. Shell seen from left side ) 



, x50. 
23. „ ,, above 



Cythere militaris. 

24. Shell seen from left side 

25. „ „ above 

26. ,, ,, below 

Cythere quadriserialis. 

27. Shell seen from left side } eA 

> x 50. 

28. „ ,, above J 

Cytheridea flavescens. 

29. Shell (female) seen from left side \ 

30. „ „ „ above ( x 6Q 

31. „ (male) ,, left side i 

32. „ „ ,, above ' 



Limnicythere Fijiana. 
„ 33. Shell seen from left side 
„ 34. ,, ,, above 

Cythere crenata. 
„ 35. Shell seen from left side 
„ 36. „ „ above 

Plate III. 

Cythere torticollis. 
Fig. 1. Shell seen from left side 
„ 2. „ „ below 

Cythere Scotti. 
„ 3. Shell seen from left side 
„ 4. „ „ above 

Cytheridea consobrina. 



1 x60. 



x60. 



x50. 



} 



x40. 



„ 5. Shell (male) seen from left side ) - 
„ 6. „ „ „ above J 

Xestoleberis tumefacta. 
„ 7. Shell seen from left side ) ™ 
,, 8. „ ,, below / 



)> >> 



524 DR G. S. BRADY ON 



Xestoleberis gracilis. 
Fig. 9. Shell seen from left side ) 
„ 10. „ „ above J 

Cytherideis baculoides. 
„ 11. Shell seen from left side ) 



, x50. 

12. ,, „ above 

Anehistrocheles fumata. 

13. Shell seen from left side 

14. „ ,, above 

Cytlieropteron rude. 

15. Shell seen from left side \ 

16. „ „ above I x 80. 

17. ,, „ front ) 



1 x50. 



Cytlieropteron longieaudatum. 
18. Shell seen from left side ) 

r 



..60 
19. „ „ above 



1 x80. 



I x80. 



Cytlieropteron coccoides. 

20. Shell seen from left side 

21. „ „ above 

Cytlieropteron trilobites. 

22. Shell seen from left side 

23. ,, ,, below 

Cytherura marcida. 

24. Shell seen from left side | ~ A 
n _ > x 60. 

25. „ „ above ) 

Cytherura entomon. 

26. Shell seen from left side 

27. „ ,, above 
27a. Another form seen from above 

Cytherura curvicostata. 

28. Shell seen from left side 

29. „ „ above 

Cytherura scutellata. 

30. Shell seen from left side ) „^ 

31. „ „ above J 

Paradoxostoma ovatum, 

32. Shell seen from left side 

33. „ „ above 

Plate IV. 



1 x80. 



} x70. 



Asterope australis. 
Fig. 1. Shell seen from left side ' ) 



. x22 

2. „ „ below 



Streptoleberis crenulata. 
„ 3. Shell seen from right side 
„ 4. „ „ above 

Sarsiella rudis. 

„ 5. Shell seen from right side i 

1 x 40. 



1 x40. 



6. „ ,, above 



} 



OSTRACODA COLLECTED IN THE SOUTH SEA ISLANDS. 



525 



Asterope cylindrica. 



Fig. 



7. Shell seen from right side 



8. 



above 



Philomedes vellicata. 
9. Shell seen from right side 

10. „ „ above 

Pleoschisma reticulata. 

11. Shell seen from left side 

12. „ ,, above 

Pleoschisma rohusta. 

13. Shell seen from left side 

14. „ „ above 

Sarsiella simplex. 

15. Shell seen from left side 

16. ,, „ above 

Bairdia ventricosa. 

17. Shell seen from left side 

18. „ ,, above 



x40. 



x40, 



x80. 



x50. 



x40. 



x60. 



Paradoxostoma Novce Caledonice. 

19. Shell seen from left side x 80. 

Paradoxostoma retusum. 

20. Shell seen from left side x 80. 

Cytherella tumida. 

21. Shell seen from right side \ 

22. „ „ above V x 80. 

23. „ „ front ; 

Loxoconcha gracilis. 

24. Shell (male) seen from left side 

25. „ ,, ,, above 

26. ,, (female) „ left side 



x60. 



Loxoconcha gibbera. 

27. Shell seen from left side 

28. ,, „ above 

Cytheropteron guttatum. 

29. Shell seen from left side 

30. „ „ above 



x60. 



x60. 



VOL. XXXV. PART II. (NO. 14). 



4 R 



Brady on South Sea Ostracoda Plate i 





13 






L6 



15 



17 



18 



21 




19 





20 




*• 



t * 




23 




22 




24 





G S.Brady, del*- 



The Srapliotone Co L* Photo Printers, Bush Hill Pipk.Enfieli 



\7/> < ■■■■ • i ; 



Brady on South Sea Ostracoda plate ji 







. ■•*««. 










v s 



■ 




^-~' 










29 






30 



\M 



31 



32 



P 




34 








G S.JSrosd?/. del*- 



The Graphotsne Co, L d PI i Bush Hill Pajk.Enfield 



Brady on South Sea Ostracoda. - plate hi 



rtfKk 



j&*£ 



f. 




Y 



i 




>sr 



9 




J't 



10 



1> 





12 
I 




13 



. 



14 






J 





G. S.J3ra,d?/. del 1 



■ iphcton. Co L d Phott. Prmtere.Bush Hill ?»**. Enfield 



r i w. 



^o N ; 



Brady on South Sea Ostracoda. - Plate iv 




13 






■ 






14 



m 



\% 



-. 





fA 





28 J^ 


?9 1MB 






■ 






•7 




6-' A' ^5/ »^/y. ^/* 



The Braphowne Co- L* PHoto Printers, Burt HfflP«.ri<, Enfield. 



( 527 ) 



XV. — On Benzyl Phosphines and their Derivatives. By Professor Letts and 
E. F. Blake, Esq., Queen's College, Belfast. (With a Plate.) 

(Read 20th May 1889.) 

CONTENTS. 



Part I. — Benzyl Phosphines. 



Introduction, 
Monobenzyl Phosphine- 



-Preparation, . 
Properties of, 
Hydriodate, . 
Hydrobromate, 
Hydrochlorate, 
Platinum Salt, 
Action of Oxygen on 
Action of Bromine on the Pro- 
ducts of the Oxidation of, . 
Action of Sulphur on, 

,, Halogens on, . 

,, Bisulphide of Carbon 
on, 



PAGE 

528 
540 
543 
544 
544 
544 
545 
545 

547 
549 
550 

551 



Monobenzyl Phosphine- 



page. 

-Action of Sulphuric Acid on, . 553 

,, ,, Chloracetic Acid on, . 553 

,, ,, Bromacetic Acid on, . 554 
,, ,, Chlorocarbonic Ether 

on, . . . 554 

Hofmann's Dibenzyl Phosphine, 554 

Isolation of Tribenzyl Phosphine, &e. , from the Product 

of the Sealed Tube Reaction, ..... 559 

Action of Haloid Compounds of Benzyl on the Primary 

Phosphine, ........ 568 

Properties of Dibenzyl Phosphine, .... 573 

,, Tribenzyl ,, 574 

Salts of Tetrabenzyl Phosphonium, .... 574 

The Bye Products of Hofmann's Sealed Tube Reaction, 576 



Part II. — Action of Alcohols on a Mixture of Phosphorus and its Iodide. 



The Methods in Use for the Preparation of Phosphines, . 589 



Part III. — The Products of the 

PAGE 

Benzyl Phosphinous Acid — Preparation and Properties, 609 

Benzyl Phosphinite of Barium, 610 

Calcium, 610 

Magnesium, .... 610 

Zinc, 611 

Cadmium, ..... 611 

Lead, ...... 611 

Copper, ..... 611 

Action of Heat on the Acid 612 

Benzyl Phosphinic Acid — Preparation and Properties, . 612 
Benzyl Phosphinate of Barium — Normal, . . . 614 
,, Acid, . . . .614 

Calcium (Normal), . . . 615 
Magnesium ,, . . .615 

Zinc ,, ... 615 

Cadmium ,, . . .616 

Lead, 616 

Copper, 616 

Silver 616 

Potassium, . . . .616 

Sodium and Ammonium, . .617 

Action of a Moderate Heat on the Acid, . . . 617 

Strong „ „ .... 618 

,, Phosphorous Acid on the Acid, . . . 618 

,, Pentachloride of Phosphorus on the Acid, . 619 

Dibenzyl Phosphinic Acid — Preparation and Properties, 619 

Dibenzyl Phosphinate of Barium, 620 

,, Calcium 621 

VOL. XXXV. PART. II. (NO. 15). 



PAGE 



The Action of Benzyl Alcohol on a Mixture of Phos- 



)xidation of Benzyl Phosphines. 




• vvv 




PAGE 


Dibenzyl Phosphinate of Magnesium, 
,, Cadmium, 






. 621 

. 621 


Copper, . 

,, Silver, . 






. 621 
. 622 


,, Sodium, . 






622 


, , Potassium, 






. 622 


,, Ammonium, 






. 622 


Action of Heat on the Acid, . 


. 




623 


,, Pentachloride of Phosphorus c 

Oxide of Tribenzyl Phosphine — Preparat 

Oxide of Tribenzyl Phosphine — Bromid* 

, , Chloride 


n the Acic 
on and Pro 
*j • 


. 623 

perties, 624 

. 624 

. 625 


,, Iodide, . 




. 625 


, , Hydrochlorate, 
,, Hydrobromate, 
,, Platinum Salt, 




626 
626 
626 


,, Palladium Salt, 




. 626 


,, Ferric Chloride Salt, 


. 626 


,, Mercuric Chloride Salt, 


626 


,, Cobaltous Chloride Salt, 


626 


,, Zinc Iodide Salt, . 


627 


,, Acetyl Chloride Com 

pound, 
,, Sulphur Compound, 
,, Nitro Compound, . 
,, Sulphonic Acid, 
,, Action of Fused Caustic 


627 
627 
627 
627 


Potash on, 


627 




4 i 


3 





528 



PROF. LETTS AND MR R. F. BLAKE ON 



PART I.— BENZYL PHOSPHINES. 



Introduction. 

The pliosphines which have been obtained as yet are not very numerous, and 
with one or two exceptions their properties have not been exhaustively studied, a fact 
which is no doubt largely due to the difficulties encountered in preparing them in 
any quantity. 

There are consequently a great many points in their history which require ex- 
amination, and the object we had in view in the present investigation was chiefly to 
extend our knowledge of this interesting group of substances. 

It may not be inadvisable, before describing the results of our experiments, to give a 
brief summary of the work done already by others. 

We discuss the methods for obtaining pliosphines in another part of this paper 
(p. 589), so that for the present we shall content ourselves with their properties and 
reactions. 

General Properties of the Primary Pliosphines. 
The following primary bases have been obtained : — 



Name. 


Condition. 


Boiling Point. 


Methyl phosphine ( 1 ), . 
Ethyl „ O, 
Iso-propyl „ ( 3 ), 
Iso-butyl „ ( 4 ), 
Iso-amyl „ ( 5 ), 
Octyl (normal) ( 6 ), 
Phenyl „ (*), 
Benzyl „ (*), 
P- Tolyl „ O, 










Gas. 
Liquid. 

» 
>> 

» 

Solid at + 4. 


- 14 
+ 25 
+ 41 
+ 62 
+ 106 
+ 184-187 
+ 160-161 
+ 180 
+ 178 



(!) Hofmann, Berichtc, iv. (1871) p. 209. 

( 3 ) Ibid. Ibid. vi. (1873) p. 292. 

( 5 ) Ibid. Ibid. vi. (1873) p. 297. 

(?) Michaelis, Ibid. vii. (1874) p. 6 and p. 1688. 

( u ) Michaelis and Panek, Annalen, 212, p. 233. 



( 2 ) Hofmann, Berichtc, iv. (1871) p. 432. 
( 4 ) Ibid. Ibid. vi. (1873) p. 296. 

( 6 ) Moslinger, Ibid. ix. (1876) p. 1005. 
( 8 ) Hofmann, Ibid. v. (1872) p. 100. 



With the exception of methyl phosphine, which is a gas, and p. tolyl phosphine, 
which is a solid, the primary bases are liquids at ordinary temperatures, insoluble in 
water, but soluble in ether, &c. Exposed to the air they fume powerfully and grow 
very hot, their vapour igniting spontaneously at times. The products of this oxidation 
appear to have been investigated only in a few cases, and chiefly in the aromatic series. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 529 

Phenyl phosphine # and tolyl phosphine t both absorb a molecule of oxygen, and are 
converted into phosphinous acids, which are monobasic, and therefore probably have the 
constitution, 

B . 

H AP = 0. 

HCK 

These acids are readily decomposed by heat, giving the primary phosphine, and the 
corresponding phosphinic acid, 

3BPH 2 2 = BPH 2 + 2RPH,0 3 , 

a reaction analogous to that which gives rise to phosphuretted hydrogen and phosphoric 
acid when hypophosphorous and phosphorous acids are heated. 

Moslinger (loc. cit.) has in part investigated the products of the spontaneous oxida- 
tion of octyl phosphine, and believes that it also is converted into a phosphinous acid. 

Submitted to the action of strong nitric acid, many of the primary phosphines (and 
probably all) absorb three atoms of oxygen, and are converted into phosphinic acids. 
This has been shown by Hofmann J to be the case in the methyl, ethyl, iso-prop}d, iso- 
butyl, and iso-amyl series. The phosphinic acids are solid substances, which with the lower 
members of the fatty series can be distilled unchanged.§ But in some other instances a 
different reaction occurs: thus phenyl phosphinic || acid, when heated slowly to 200° C, 
gives a pyro acid, while when rapidly heated to 250° C. it decomposes into benzol and 
meta-phosphoric acid, 

C 6 H 5 PH 2 3 = C 6 H c +HP0 3 . 

The phosphinic acids are all dibasic, and no doubt have the structure — 

B , 

HO->PO . 
HCK 

All the primary phosphines have distinct alkaline properties. They combine readily 
with hydracids, forming crystalline compounds which can as a rule be volatilised (with dis- 
sociation more or less complete), and which resemble the compounds of phosphuretted 
hydrogen in other respects, particularly in being instantly decomposed by water with 
liberation of the phosphine. The hydrochl orates combine with chloride of platinum to 
give chloroplatinates. The salts of the primary bases with oxyacids have been scarcely at 
all investigated. 

The action of halogens on mono-phosphines has not been sufficiently investigated. 
Methyl and ethyl phosphine take fire when they come in contact with chlorine or 
bromine,! but in other cases the action does not appear to have been studied. 

* Michaelis, Berichte, x. (1877) p. 807. + Michaelis and Panek, Annalen, 212, p. 234. 

J Hofmann, Berichte, v. (1873) p. 110. § Hofmann, Berichte, vi. (1873) p. 303. 

|| Michaelis, Berichte, vii. (1874) p. 1070. IF Hofmann, Berichte, iv. (1871) pp. 433 and 609. 



530 PROF. LETTS AND MR R. F. BLAKE ON 

Sulphur acts on the primary bases. With the methyl and ethyl derivatives, com- 
pounds have been obtained but not investigated.* With phenyl phosphine sulphur acts 
slowly in the cold, rapidly at a high temperature.t Two substances are produced, one, 
a thick liquid soluble in ether, having the composition (C 6 H 5 )PH 2 S, the other a crystalline 
product to which Michaelis assigns the formula (C 6 H 5 P) 3 S. The first of these bodies 
decomposes when heated in the following manner : — ■ 

2C C H 5 PH 2 S = C G H 5 PS + C 6 H 5 PH 2 + H 2 S . 

In view of the analogies existing between nitrogen and phosphorus, considerable 

interest is attached to the action of carbonyl chloride and bisulphide of carbon on the 

primary phosphines. Bisulphide of carbon acts upon both methyl and ethyl phosphine,| 

but the products have not been investigated. Michaelis and Dittler § have studied 

the action of both reagents on phenyl phosphine. When carbonyl chloride is 

passed slowly into that substance, an energetic reaction occurs in the following 

manner : — 

2CO Cl 2 + C e H 5 PH 2 = C e H 5 PCl 2 + 2CO + 2HC1 . 

Phenyl phosphine and bisulphide of carbon act upon each other when heated in a sealed 
tube at 150°, and sulphuretted hydrogen is liberated. The product of the reaction is a 
resinous body (C 6 H 5 PHCS) 2 S, and the reaction itself proceeds according to the equa- 
tion, 

2C C H 5 PH 2 +2CS 2 = (C 6 H 5 PHCS) 2 S + H 2 S . 

Michaelis and Dittler were not successful in their attempts to prepare a phosphorised 
mustard oil from this compound. 

They were equally unsuccessful in obtaining a phosphorised carbylamine by the 
action of chloroform and caustic potash on phenyl phosphine. It is true that a reaction 
occurs, but its course is completely different from that which takes place with an 
amine, viz., 

C„H 5 PH 2 + 4KHO + CHCI3 = C c H 5 PHK0 2 + 3KC1 + CH 3 OH + H 2 . 

It thus appears that, in their behaviour with carbonyl chloride and a mixture of caustic 
potash and chloroform, primary phosphines behave in an entirely different manner from 
the corresponding amines, though there is a certain degree of analogy as regards the 
action of both on bisulphide of carbon. This difference is no doubt due to the strong 
affinity of phosphorus for electronegative elements, such as the halogens, oxygen, and 
sulphur. 

Primary phosphines readily combine with alkyl iodides to give hydriodates of 
secondary phosphines. 

♦Hofmann, Berichte, iv. (1871) pp. 433 and 610. + Michaelis, Berichte, x. (1877) p. 810. 

J Hofmaxx, Berichte, iv. (1871) pp. 433 and 610. § Michaelis and Dittler, Berichte, xii. (1879) p. 338. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 



531 



General Properties of the Secondary Phosphines. 
The following secondary phosphines have been obtained : — 



Name. 



Condition. 



Boiling Point. 



Dimethyl 
Di-ethyl 

Di-isopropyl 

Di-isobutyl 

Di-isoamyl 

Di-phenyl 

Methyl-isopropyl 

Iso-propyl, isobutyl 



phosphine ( x ), 

„ o, 

„ ( 3 ), 

„ <\) 

( 5 )> 

( e ), 

CO. 

( 8 ), 



Liquid. 



25 

85 

118 

153 

210-215 

280 (about) 

78-80 

139-140 



( x ) Hofmann, Berichte, iv. (1871) p. 610. ( 2 ) Hofmann, Berichte, iv. (1871) p. 433. , 

( 3 ) Ibid. Ibid. vi. (1873) p. 294. ( 4 ) Ibid. Ibid. vi. (1873) p. 296. 

( 5 ) Ibid. Ibid. vi. „ p. 298. ( 6 ) Michaelis, Ibid. xv. (1882) p. 80lA. 

C) Ibid. Ibid. vi. „ p. 295. ( 8 ) Hofmann, Ibid. vi. (1873) p. 300. 

* Dibenzyl phosphine is not mentioned, for reasons which will become apparent in the body of the paper. 

All the secondary phosphines obtained as yet are liquids, having a powerful odour. 
They are insoluble in water, but soluble in ether, &c. 

They have as a rule a strong attraction for oxygen, fuming and growing hot on 
exposure to the air, and often inflaming spontaneously. In some cases they appear to 
have even a greater affinity for oxygen than the primary bases. This is so according to 
Hofmann with all the secondary phosphines of the fatty series which he obtained. But 
apparently it is not the case with diphenyl phosphine. The products of this spontaneous 
oxidation do not appear to have been examined. But the products of their oxidation by 
nitric acid have been investigated, chiefly by Hofmann.* They are in all cases phosphinic 
acids, R 2 PH0 2 , which are monobasic, and no doubt have the constitution, 

R 






HO 

These acids are probably also produced when the chlorides E 2 PC1 are oxidised by 
nitric acid. Such is at least the case with (C G H 5 ) 2 PC1. Some of them can be distilled 
unchanged, e.g., dimethyl phosphinic acid; others, e.g., diphenyl phosphinic acid, lose 
water, and give pyro acids. 

Secondary phosphines combine with acids, the resulting salts being far more stable 
than those of the primary bases. Thus in most cases they are not decomposed by 
water, though some are (e.g., salts of diphenyl phosphine). Comparatively little is 
known regarding secondary phosphines, and very few of their compounds have been 
investigated. 

Both sulphur and bisulphide of carbon act upon them, but the products have not 



* Hofmann, Berichte, v. (1872) p. 104, and vi. (1873) p. 303. 



532 



PROF. LETTS AND MR R. F. BLAKE ON 



been investigated. They readily combine with alkyl iodides, giving hydriodates of 
tertiary phosphines. 

General Properties of Tertiary Phosphines. 

The following tertiary bases have been obtained : — 



Name. 


Condition. 


Boiling Point. 


Trimethyl phosphine ( 1 ), 


Liquid. 


40-42 


Tri-ethyl 


( 2 ), 




)> 


127 


Tri-isopropyl „ 


( 3 )> 




>> 


? 


Tri-isobutyl „ 


( 4 ), 




» 


215 


Tri-isoamyl „ 


( 5 ), 




)> 


about 300 


Tri-phenyl „ 


( 6 )> 




Solid. 


above 360 


Ethyl-isopropyl-isobutyl „ 


0. 




Liquid. 


about 190 


Methyl-diphenyl „ 


( 8 ), 




>» 


284 


Ethyl-diphenyl „ 


<•>, 




j» 


293 


Di-ethyl-phenyl „ 


( 10 )> 




>> 


220 


Di-methyl-ethyl „ 


< u ). 




» 


83-85 


Di-ethyl-methyl „ 


( 12 >, 




» 


110-112 


Di-ethyl-propyl „ 


e 3 ), 




j> 


146-149 


Di-ethyl-isoamyl „ 


( u ), 




» 


185-187 


Di-ethyl-benzyl „ 


( 15 ), . 




» 


252-255 


Ethyl-dibenzyl „ 


( 16 ;, . 




» 


320-330 


Dimethyl p. tolyl „ 


( 17 ), 




» 


210 


Di-ethyl p. tolyl „ 


( 1S ), • 




>! 


240 


Dimethyl xylyl „ 


o, . 




» 


230 


Diethyl xylyl 


( 20 )> • 


» 


260 



( J ), ( 2 ) Hofmann and Cahotjrs, Amu de Chim. unci dc Phys., 3, vol. li. p. 35. 

( 3 ), ( 4 ). ( 5 ) Hofmann, BericMe, vi. (1873) pp. 292 and 304. ( 6 ) Michaelis, Beriehte, xv. (1882) p. 801a. 

( 7 ) Hofmann, Beriehte, vi. (1873) p. 304. ( 3 ) Ibid. Annalen, 181, p. 345. 

( 9 ), ( 10 ) Michaelis and Link, Annalen, 107, p. 210. 

( n ). ( 12 ). ( 13 ), ( 14 )» ( 16 ), ( 16 ) Collie, Chem. Soc. Jour., 1888, p. 714. 

( 17 ), ( 18 ). ( 19 ), ( 20 ) Czimatis, Beriehte, xv. (1882) 2014. 

The tertiary phosphines hitherto obtained are, with the exception of triphenyl 
phosphine, liquids at ordinary temperatures, having a powerful odour. They are 
insoluble in water, but soluble in ether, &c. As a rule, they oxidise rapidly in contact 
with the air, fuming and growing hot, and in some cases igniting spontaneously. The 
product of this oxidation is a tertiary phosphine oxide of the formula R 3 PO, and no 
doubt of the constitution 

E-)P=0. 

The final products of the oxidation of phosphuretted hydrogen, and of primary, 
secondary, and tertiary phosphines are therefore respectively — 

H 3 P0 4 
RH 2 P0 3 
R 2 HP0 2 
R 3 PO. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 533 

The amount of oxygen absorbed by the phosphine decreasing in a regular manner as 
the series is ascended. 

The oxides of tertiary phosphines are solid substances of remarkable stability. They 
can in the majority of cases be distilled, and even boiled with nitric acid without change. 
By no means as yet discovered can they be reduced. Hydracids combine with them, 
and they give crystalline compounds, with a number of metallic salts, such as the chlorides 
of platinum, zinc, mercury, iron, cobalt, &c, also in some cases with chloride of acetyl, 
bromine, and sulphur. 

Tertiary phosphines also combine with the elements of the sulphur group, forming 
compounds analogous to the oxides. 

The salts of tertiary phosphines are readily obtained by dissolving the bases in acids. 
They are stable, and are not, as a rule, decomposed by water. Their compounds with 
hydracids have been chiefly studied; those containing oxyacids have not been investi- 
gated (with very few exceptions). The haloid salts dissociate to a greater or less extent 
on heating. Their hydrochlorates combine with chloride of platinum to give chloro- 
platinates of normal composition. 

Some of the tertiary phosphines combine with chloracetic acid to give hydrochlorates 
of phosphorised betaines. At present only two or three of these substances have been 
obtained — tri-methyl phosphorus betaine by Meyer,* the corresponding ethyl derivative 
by HoFMANN,t and in addition to these two the closely allied compound, tri-methyl phos- 
phorus benzo-betaine hydrochlorate, by Michaelis and Czimatis,| 

(CH 3 ) 3 PV^_ COOH . 

The compounds of these phosphorised betaines are stable and well-defined substances. 
One of us§ has investigated the reactions and decompositions of the ethylated body, 
which are of some interest. 

The hydrate and the salts of this betaine lose carbonic anhydride when heated, and 
give rise to the hydrate or salt of methyl-tri-ethyl-phosphonium, 

(C 2 H 5 ) 3 P< CH2 _ COOH = C0 2 + (C 2 H 5 ) 3 P<^ CH ^ 

a reaction which is entirely analogous to that occurring when the corresponding 
sulphur compounds (thetines) are heated, 

(CH 3 ) 2 S^ CH2 _ COOH =C0 2 + (CH 3 ) 2 S^ CH3 

While it is perfectly different from that which the true (nitrogen) betaines experience, as 
they either dissociate into the original trialkyl-amine and the group X - CH 2 — COOH 
(or the products of its decomposition), or distil unchanged. || 

* Meyer, Berichte, iv. (1871) p. 734. + Hofmann, Proc. Boy. Soc, xi. p. 530. 

X Michaelis and Czimatis, Berichte, xv. (1882) p. 2018. § Letts, these Transactions, xxx. part 1, p. 285. 

|| Bruhl, Annalen, 177, p. 214. 



534 PROF. LETTS AND MR R. F. BLAKE ON 

Treated with caustic potash, all the salts of tri-ethyl phosphorus betaine yield tri- 
ethyl phosphine oxide, 

(C 2 H 5 ) 3 P<(g H _ COOH + 2KHO = (C 2 H 5 ) 3 PO + KX+H 2 + CH 3 -COOK . 

Several of the tertiary phosphines combine directly and energetically with a molecule 
of bisulphide of carbon to give highly characteristic compounds, usually of a red colour, 
and possibly having the constitution, 

■0 



R 3 P< 



So characteristic and so readily formed is this compound in the case of tri-ethyl phos- 
phine, that its production may be employed as a test either for bisulphide of carbon or 
for the phosphine itself. As yet these (bisulphide) compounds have been obtained only 
with methyl, ethyl, and iso-propyl phosphine, and with those of the aromatic phos- 
phines containing ethyl or methyl groups. 

According to Czimatis, # these mixed phosphines combine very easily with bisulphide 
of carbon if they contain methyl, the readiness with which combination occurs diminish- 
ing however, in proportion to the molecular weight of the aromatic radical, while, if they 
contain ethyl, combination occurs only slowly and with difficulty. HoFMANNt has 
somewhat exhaustively studied the compound of tri-ethyl phosphine and the bisulphide, 
which forms with explosive violence. Among its properties are the following : — It is 
insoluble in water, difficultly soluble in ether, but easily dissolves in hot alcohol, from 
which it separates on cooling in red needles like chromic anhydride. From an ethereal 
solution it is deposited by spontaneous evaporation in large deep red monoclinic crystals 
exhibiting dichroism, which melt at 95° and volatilise at 100°. It is soluble in strong 
hydrochloric acid, and if the solution is mixed with platinic chloride, a yellow amorphous 
compound is produced, 2(C 2 H 5 ) 3 PCS2,PtCl 4 . With silver oxide or nitrate, it is decomposed 

as follows : — 

(C 2 H 5 ) 3 PCS 2 +2Ag 2 = Ag 2 S+Ag 2 + C0 2 +(C 2 H 5 ) 3 PS ) 

and moist air produces a similar change. But if it is heated with water to 100° C, the 
following reaction occurs : — 

4(C 2 H 5 ) 3 PCS 2 + 2H 2 = 2(C 2 H 5 ) 3 PS + (0 2 H 6 ) 3 PO + (C 2 H 5 ) 3 (CH 3 )POH + 3CS 2 . 

Heated with sulphuretted hydrogen, it suffers the following change — 

3(C 2 H 5 ) 3 PCS 2 + H 2 S = 2(C 2 H 5 ) 3 PS + (CH 2 S)(C 2 H 5 ) 3 PCS 2 + CS 2 . 

The action of halogens upon tertiary phosphines has not been very fully studied. 
Probably direct addition would occur in all cases. This has been proved to take place 
with tri-ethyl phosphine if the halogen is allowed to act very gradually upon it. The 
chloride (C 2 H 5 ) 3 PC1 2 thus obtained is crystalline, melting at 100° and volatilising readily, 

* CV.imatis Berichte, xv. (1882) p. 2016 t Hofmann, Phil. Trans., I860, p. 431. 



BENZYL PHOSPHTNES AND THEIR DERIVATIVES. 535 

though its boiling point is high. Similar compounds of bromine and iodine have been 
obtained. 

Compounds of tri-methyl and tri-ethyl phosphine with mustard oils, are formed 
easily, and give crystalline hydrochlorates. They no doubt have the constitution, 

S = C = <^E S 

[_Note. — Some of the aromatic tertiary phosphines, especially tri-phenyl phosphine, 
have properties which differ materially from those of other tertiary phosphines. Thus 
tri-phenyl phosphine is a crystalline solid having scarcely any odour, and it does not 
oxidise spontaneously. It is remarkably stable, and is not attacked by chlorine even 
when heated. The hydriodate and hydrochlorate are formed when it is dissolved in the 
warm concentrated hydracids, and are crystalline, but on adding water they dissociate. 
By treating the phosphine with bromine and an alkali, or by oxidising it with hydrochloric 
acid and chlorate of potash, the hydrate (C 6 H 5 ) 3 P(OH) 2 is obtained as a crystalline 
solid. This when heated to 100° readily loses water, and is converted into the oxide, a 
substance which is not acted upon by bromine, oxygen, sulphur, &c. By dissolving the 
phosphine in fuming nitric acid a nitrate of the formula (C 6 H 5 ) 3 P(N0 3 ) 2 is obtained.] 

Tertiary phosphines, apparently without exception, unite with alkyl iodides to form 
phosphonium salts. 

General Properties of Quaternary Compounds (Phosphonium Salts). 

So many of these bodies have been obtained that a list appears inadvisable. It would 
include derivatives of the series C n H 2n+1 to the 5th term, one or two of the series 
C n H 2n _7, and a large number of mixed phosphoniums containing various radicals, among 
which are vinyl, allyl, and ethylene. 

The phosphonium salts are the most stable of all organic phosphorus compounds. 
None are decomposed by water, and most of them can be obtained readily in the crystal- 
line state by evaporating their solutions. 

As a rule, they are soluble in water and in alcohol. They are readily prepared from 
their iodides, either by double decomposition with a silver salt, or by first obtaining 
their hydrates (by the action of hydrate of silver), and subsequently neutralising the 
solution with the acid. 

The hydrates EJPOH are solid substances, having a powerful alkaline reaction and 
many properties similar to those of an alkali. Indeed, in the case of tetrethyl phos- 
phonium hydrate, the only remarkable point of difference between it and caustic potash 
(so far as its reactions with metallic salts, &c, are concerned) is that when added to a 
zinc or aluminum salt, the zinc or aluminum hydrate, which is at first precipitated, is 
insoluble in an excess. Phosphonium hydrates are decomposed when heated, and in 
some cases, when their solutions are boiled or at the moment of production, into a 
tertiary phosphine oxide and a hydro-carbon, 

R 4 POH = R 3 PO + R-H. 

VOL. XXXV. PART II. (NO. 15). 4 T 



536 PROF. LETTS AND MR R. F. BLAKE ON 

The action of heat upon the salts of the phosphoniums has been investigated in a 
number of cases, partly by one of us * and N. Collie, and partly by the latter chemist 
alone. 

As regards the haloid salts, the chlorides decompose almost quantitatively into a 
hydrocarbon and a tertiary phosphine hydrochlorate (Collie), furnishing an excellent 
method for " retrograding" from quaternary to tertiary bodies. 

" When the phosphonium chloride contains several ethyl groups, then if more than 
one of the latter is present, ethylene is always formed, e.g., 

(C 2 H 5 ) 3 (C 7 H 7 )PC1 - (C 2 H 5 ) 2 (C 7 H 7 )P.HC1 + C 2 H 4 . 

But when only one ethyl group is present, then, although ethylene is still formed, two 
decompositions occur, e.g., 

(1) 2(C 2 H 5 )(CH 3 )3PC1 = 2(C 2 H 5 )(CH 3 ) 2 P.HCH-C 2 H 4 

(2) (C 2 H 5 )(CH 3 ) 3 PC1= (CH 3 ) 3 P.HC1 + C 2 H 4 . 

If we compare the decomposition by heat of phosphonium chlorides with the 
decomposition of any of the compound ammonium salts, it must be with the hydroxides 
and not wth the corresponding chlorides" (CoLLiE),t e.g., 

(C 2 H 6 ) 4 PC1 =(C 2 H 5 ) 3 P.HC1 + C 2 H 4 
(C 2 H 5 ) 4 NOH = (C 2 H 5 ) 3 N + C 2 H 4 +H 2 

(C 2 H 5 ) 3 (C 7 H 7 )P(OH) - (C 2 H 5 ) 3 PO + C 7 H 8 
(C 2 H 5 ) 3 (CH 3 )NC1 =(C 2 H 5 ) 3 N +CH 3 C1. 

The effect of heat on phosphonium salts derived from oxy-acids is completely 
different. In the case of the ethyl series at all events, they suffer, as a rule, at least 
two and occasionally three different and distinct decompositions. In one of these the 
molecule splits up into three new groups, consisting respectively of carbonic anhydride, 
a (paraffin) hydrocarbon, and the tertiary phosphine. In the other, two hydrocarbons are 
formed, namely, an define and a paraffin, in addition to carbonic anhydride and the 
tertiary phosphine. Whilst in the third, a totally different change occurs, in which 
only two products are formed, namely, the oxide of the tertiary phosphine and a ketone, 

(1) Et 3 P<^ R =Et 3 P + C0 2 +C 2 H 5 R. 

(2) Et 3 P<^^ = Et 3 P+C0 2 +C 2 H 4 + RH. 

(3) Et 3 P<g^, R = Et 3 PO + C 2 H 5 .CO.R . 

It is possible, if not indeed probable, that the third reaction occurs subsequently to 
the first, and that it really depends upon the reducing action of the triethylphosphine 

* Letts and Collie, (1) these Transactions, xxx. part 1, p. 213; and (2) Phil. Mag., August 1886. 
t Collie, (1) Chem. Soc. Jour., 1888, p. 636; aud (2) Ibid., p. 714. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 537 

upon the carbonic anhydride, at the high temperature at which the decomposition usually 
occurs, whereby carbonic oxide is liberated, which combines with the hydrocarbon radical 
in statu nascendi, forming a ketone : — 

Et 3 P+C0 2 +(Et)+(R) = Et 3 PO + EtCOR 
(or Et 3 P+(OCO-R) + (Et) = Et 3 PO + EtCOR). 

If we merely consider the third kind of decomposition alone, it appears to be, to a 
certain extent, analogous to the decomposition which a sulphine compound suffers when 
heated, the difference depending on the greater attraction which phosphorus has for 
oxygen, compared with that of sulphur for the same element. In both cases a hydrocarbon 
group is detached from the molecule, and also the residue of the acid, but while with the 
sulphur compounds these two simply combine (forming a compound ether), and leave a 
hydrocarbon sulphide, in the case of the phosphonium salt the acid residue is reduced by 
the tertiary phosphine, and the group thus left combines with the hydrocarbon radical, 
forming a ketone. 

Thus— 

Q t OCR =Et 2 S + EtOOCR. 



Et 2 S< 

Et 3 P<g : Q CR =Et 3 PO + EtOCR 



A result of this kind is in perfect harmony with the views already expressed by 
Crum Brown and Letts* regarding the analogies and differences existing between 
phosphorus and sulphur and their compounds. 

When the phosphonium salts contain ethylene they suffer a different decomposition 
under the influence of heat ; at least this has been ascertained to be the case with the 
bromide of bromo-ethylene-triethylphosphonium, and the bromide of hydroxy-ethylene- 
triethyl phosphonium, which decompose as follows :t — 

C 2 H 4\pJq H ) B r =HBr+(C 2 H 3 )(C 2 H 5 ) 3 PBr 

and also in that of the hydrate of ethylene-hexethyl diphosphonium, which decomposes 
according to the equation, 

Masson and Kirkland J have studied the action of bromine and chlorine on the salts of 
tetrethyl phosphonium, the results showing a very close similarity between the poly-haloid 
derivatives of tetrethyl phosphonium and those of trimethyl sulphine and of tetramethyl 
ammonium previously described by Dobbin and Masson. § The tendency to form solid 

* Crum Brown and Letts, these Transactions, xxviii. p. 371 ; Letts, these Transactions, xxx. p. 285. 

t Hofmann. t Masson and Kirkland, Chem. Soc. Jour., 1889 (Trans.), p. 126. 

§ Dobbin and Masson, Chem. Soc. Jour., 1885 (Trans.), p. 56, and 1886, p. 846. 



538 



PROF. LETTS AND MR R. F. BLAKE ON 



poly-haloid compounds is, however, more marked. They give the following table, contain- 
ing a list of the new substances, the methods of forming them, and their chief properties : — 



No. 


Substance. 


Method of Formation. 


Temperature of 
Formation. 


Chief Properties. 


1 


PEt 4 IBr 4 ? 


Br 2 on iodide. 


cold. 


Bright red crystals, solid ; unstable. 


2 


PEtJBr 


Alcohol on I. 




Orange crystals ; stable. 


3 


PEt 4 ICl 4 


Cl 2 on iodide. 


70° 


Yellow crystal mass ; unstable. 


4 


PEt 4 ICl 2 


Alcohol on 3. 




Yellow crystals ; stable. 


5 


PEt 4 Br 7 


Br 2 on bromide. 


110 


Yellow crystal mass ; unstable. 


6 


PEt 4 Br 3 


Alcohol on 5 or 9. 




Red crystals ; stable. 


7 


PEt 4 Cl 3 


Cl 2 on chloride. 


110 


Yellow crystal mass ; unstable. 


8 


(PEt 4 ) 2 S0 4 Br.„ ? 


Br 2 on sulphate. 


cold. 


Red, liquid ; unstable. 


9 


(PEt 4 ) 2 S0 4 Br 12 


Br 2 on sulphate. 


110 


Red, solid ; unstable. 


10 


(PEt 4 ) 2 S0 4 Cl 4 


Cl 2 on sulphate. 


130 


Yellow, solid ; unstable. 



Organic Phosphorus Compounds which cannot be placed in any of the above Groups. 

A phosphorised kakodyl (CH 3 ) 4 P 2 (the methyl analogue of liquid phosphuretted 
hydrogen) was obtained by Thenard # by the action of chloride of methyl on phosphide 
of calcium, and is interesting not alone as being the sole representative (as yet prepared) 
of its class, but also as having been probably the first phosphine obtained. Thenard 
describes it as a colourless, highly refractive liquid, of an odour recalling that of kakodyl 
itself, insoluble in water, and boiling at about 250°. It inflames spontaneously in 
contact with air, but if oxidised slowly, gives a crystalline acid (CH 3 ) 4 P 2 H 2 4 := 
(CH 3 ) 2 PH0 2 (dimethyl phosphinic acid), analogous to kakodylic acid. Treated with an 
excess of hydrochloric acid, it is converted into trimethyl phosphine, and a solid yellow 
substance (CH 3 ) 2 P 4 (which is also formed in the original reaction), and which Thenard 
regarded as the methyl analogue of solid phosphide of hydrogen. 

Michaelis t obtained a substance, which he named diphosphenyl, or phospho-benzol, 
C 6 H 5 — P = P — C 6 H 5 (corresponding to azobenzol), by the action of phenyl phosphorous 
chloride on monophenyl phosphine, 

C 6 H 5 PC1 2 +C 6 H 5 PH 2 = (C 6 H 5 ) 2 P 2 + 2HC1. 
It is a pale yellow powder, insoluble in water, alcohol, and ether, but readily soluble in 
hot benzol, and is slowly oxidised by the air to (C 6 H 5 ) 2 P 2 0. Treated with chlorine, 
phenyl phosphorous chloride is regenerated. With nitric acid it is oxidised to phenyl 
phosphinous acid C 6 H 5 PH 2 2 , if the acid is dilute, but to phenyl phosphinic acid, 
C 6 H 5 PH 2 3 , if the acid is strong. Treated with hydrochloric acid, it reacts so as to 
regenerate the substances from which it is formed. 

Michaelis | also obtained a substance, which he called di-phospho-benzene hydrate, 
C 6 H 5 — P = P — OH, by the action of spontaneously inflammable phosphuretted hydrogen 
on phenyl phosphorous chloride. It is a yellow powder, soluble with ease in bisulphide of 
carbon, taking fire on exposure to air, and oxidised by nitric acid to phenyl phosphinic 

* Thenard, Comptes Rendus, xxi. p. 144, and xxv. p. 829. 

t Michaelis, Berichte, x. p. 807. J Michaelis, Berichte, vii. (1875) p. 499. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 539 

and phosphoric acids. In addition to the above, Michaelis * obtained a phenylated solid 
phosphide of hydrogen, (C 6 H 5 )HP 4 , by treating phenyl phosphorous chloride with a quantity 
of water insufficient for complete decomposition (for instance, by keeping it in a badly- 
stoppered bottle). It is a dark yellow amorphous body, having a faint odour of phenyl 
phosphine, soluble in hot bisulphide of carbon, but insoluble in water, alcohol, and ether. 
Treated with chlorine, it reacts as follows : — 

(C 6 H 5 )HP 4 + 6C1 2 = 3PC1 3 + (C 6 H 5 )PC1 2 + HC1 . 

Nitric acid oxidises it to a mixture of phenyl phosphinic and phosphoric acids. 

The action of phosphonium iodide on aldehydes has been studied by GiRARD,t while 
that of phosphuretted hydrogen and hydrochloric acid on the same bodies and on ketonic 
acids has been investigated by Messinger and Engels. J Girard obtained products of 
addition containing four molecules of the aldehyde (valeric, propionic, salicylic, and benzoic) 
to one of phosphonium iodide. 

Messinger and Engels obtained similar bodies by acting upon the aldehydes with 
hydrochloric acid and phosphuretted hydrogen. The compounds thus formed are for the 
greater part solid, crystalline, and fairly stable. By treatment with water, they are 
decomposed, and the aqueous solution gives the reactions of hydrochloric acid and 
phosphuretted hydrogen. Their constitution is probably represented by the formula 
(R-CHOH) 4 PCl. Chloride of tetra-hydroxyethylidene phosphine (C 2 H 5 0) 4 PC1, is 
decomposed by caustic potash into the free phosphine (C 2 H 5 0) 3 PC 2 H 4 0, and the hydrate 
(C 2 H 6 0) 4 P(OH). Benzaldehyde and its mono-nitro derivative give compounds which 
differ from those obtained in the fatty series, in that they contain no hydracid. 

Messinger and Engels have summarised the result of their researches as follows : — 
(1) Phosphuretted hydrogen does not act on an aldehyde alone, but is absorbed if at the 
same time a hydracid is present. The absorption occurs more completely if the aldehyde 
is largely diluted with ether. (2) The aldehydes of the fatty series combine with a molecule 
of phosphuretted hydrogen and a molecule of hydracid, while those of the aromatic series 
combine with phosphuretted hydrogen only, though in order that the compound shall be 
formed the presence of the hydracid is necessary. (With benzoic aldehyde the compound 
has the formula (C 6 H 5 COH) 4 PH 3 ). (3) The phosphorised derivatives of the fatty 
series have an unpleasant smell, and are decomposed by water, while those of the 
aromatic series have no odour, and are nearly insoluble in water. All are soluble with 
difficulty in ether, and in some cases insoluble. 

By the substitution of a ketonic acid for an aldehyde in the above reaction, com- 
pounds are produced in certain cases. Thus lsevulinic acid gives an oil and pyruvic acid 
a solid compound either — 


(CH 3 - CO - CO') 3 P or(CH 3 - (Cb = o) 3 P , 

* Michaelis, Berichte, xi. (1878) p. 885. 

t Girard, Ann. de Chim. und de Phys. [vii.], ii. p. 50. 

% Messinger and Engels, Berichte, xxi. (1888) p. 328a and p. 2919a. 



f)40 PROF. LETTS AND MR R. F. BLAKE ON 

which is a well-defined crystalline body, having neither basic nor acid properties, soluble 
in alkalies with decomposition, and also decomposed when heated with acids. It 
dissolves, however, without change in glacial acetic acid, and crystallises out on cooling. 
Boiled with water, it is decomposed into the substances from which it was originally 
produced. It forms crystalline compounds with aniline, phenyl hydrazine, and toluene 
diamine. 

On reading over the above, it will be seen that, although a fair amount of work has 
been done with the phosphines, they have by no means been exhaustively studied, and 
that many of their properties and reactions remain to be investigated. It was our wish 
to fill up some of these gaps in their history, and we chose the benzyl phosphines for 
investigation for several reasons, among which was the fact that the primary 
phosphine is a liquid at ordinary temperatures, also on account of the well-known 
chemical activity of the benzyl compounds, and partly because one of us and N. Collie 
had already studied somewhat exhaustively the compounds of tetrabenzyl phosphonium. 
Hofmann* was the first to obtain monobenzyl phosphine, and as he believed di- 
benzyl phosphine also, by heating a mixture of chloride of benzyl, phosphonium iodide, 
and oxide of zinc in sealed tubes. He apparently submitted the two substances to a 
somewhat cursory examination, and only determined their leading properties. He 
mentions that bye products are formed, but these he did not investigate. 

We have repeated Hofmann's experiments, and have submitted both the primary 
phosphine and also the substance which he regarded as the secondary derivative to a 
very careful examination. We have also isolated the other products and bye products 
of the reaction, and have determined their composition, and as far as possible their pro- 
perties also. In the course of the research we incidentally discovered a method for 
preparing all the products of the oxidation of the benzyl phosphines. We describe this 
method, and also the properties of the oxidised substances. 



Monobenzyl Phosphine and its Derivatives. 

Preparation of Monobenzyl Phosphine. — Hofmann recommends digestion during six 
hours at 160° C. of a mixture of 4 parts of oxide of zinc, 16 of iodide of phosphonium, 
and 12 of chloride of benzyl. Experiments conducted in this way with commercial 
chloride of benzyl from Kahlbaum gave in the tubes a viscous semicrystalline mass. To 
obtain a good result, thorough mixing of the materials in the sealed tubes by shaking 
before heating seemed to be necessary. On opening the tubes much phosphuretted 
hydrogen escaped, but on heating for a longer period or to a higher temperature, the 
escaping gas seemed to consist of hydrochloric acid only. It was soon found that at the 
temperature of 160° O, a great deal of hydrochloric acid is formed, and but little of the 
primary phosphine. The best results were obtained by a six hours' digestion of the 

* Hofmann, Berichte, v. (1872) p. 100. 



BENZYL PHOSPHLNES AND THEIR DERIVATIVES. 541 

mixture at a temperature of 120° C. Experiments tried at 100° to 110° C. showed that 
but little of the primary phosphine is formed. 

With the quantities Hofmann recommends and a digestion for six hours at 120°, the 
tubes when cold contain a viscous semitransparent mass, sometimes of a brown colour 
sometimes red and opaque from the separation of free phosphorus. Above this a small 
quantity of a liquid usually floats, which at times is mobile, but at others thick and 
slightly fluorescent. A few crystals of undecomposed iodide of phosphonium are also 
frequently present. If pure chloride of benzyl is used there is usually no liquid floating 
on the surface of the product. 

After opening the tubes, their contents were transferred to a flask, and submitted 
to the action of steam and water by means of the apparatus shown in figure 1 of the 
Plate. 

A is a wide-mouthed flask provided with a large cork, with a large and somewhat 
conical hole, through which the inverted tube B passes. 

C is a bent tube, which can be raised or lowered at will according to the length of B, 
so that its upper extremity may be near the closed end (of B). Through this tube steam 
from D, or carbonic acid from E, can be passed. 

Two other tubes, F and G, pass through the same cork — one for blowing steam 
through the product, the other for carrying volatilised substances into the condenser H. 

The apparatus was employed as follows: — After opening the tubes containing the 
product of the reaction, their tops were cut off with a file, and they were then placed in 
front of a strong fire to liquefy their contents and to drive off as much hydrochloric and 
hydriodic acids as possible. They were then allowed to cool in a nearly horizontal 
position, so that their contents solidified on one side of the tube, leaving a clear passage 
along their whole length. Next a stream of carbonic anhydride was passed through F 
into A until it was quite full of the gas. Steam was then passed through C, and one of 
the tubes B inverted over its open end, and rapidly pushed into the cork. Its contents 
almost immediately liquefied and dropped into A without coming in contact with the air. 
When the tube was completely washed out, it was removed and another one substituted. 
The contents of twelve tubes were usually worked up at one operation (each containing 
4 grms. oxide of zinc, 16 grms. phosphonium iodide, and 12 grms. chloride of benzyl) — 
the last tube being left in the position shown in the figure. The tubes F and C were 
then interchanged, so that steam passed through F, while a gentle current of carbonic 
anhydride was passed through C. 

By this means the product was submitted to the prolonged action of steam and water 
in a current of carbonic anhydride. The primary phosphine thus set free passed over 
along with the steam, and was condensed in H, and then flowed into the separating funnel 
I, which contained common salt to give greater density to the water, and thus cause the 
phosphine to rise to its surface, as its specific gravity is almost the same as that of water. 

Without the above apparatus, we found it is almost impossible to separate the 
phosphine without considerable loss through oxidation. 



542 PROF. LETTS AND MR R. F. BLAKE ON 

Under favourable circumstances, from 60-70 grms. of crude monobenzyl phosphine 
were obtained from 360 grms. of benzyl chloride ; or, in other words, each tube yielded 
a little over 2 grms., consequently the preparation of even the crude phosphine in any 
quantity is extremely tedious and troublesome. 

The residue in the flask A always contained in addition to water, a brown viscous and 
insoluble liquid, which as a rule solidified on cooling. Both it and the water contained 
phosphorised benzyl derivatives. We shall first discuss the properties of the primary 
phosphine, and afterwards the nature of the various bye products, and the methods we 
employed for isolating them. 

Monobenzyl Phosphine. — Hofmann purified the crude phosphine by fractional distilla- 
tion only. He states that after two rectifications in a stream of hydrogen it is obtained of 
the constant boiling point 180°. 

Our own experiments, repeated again and again, and with the greatest care, have 
satisfied us that the pure phosphine cannot be readily obtained thus. We have at 
different times operated upon two to three hundred grams, and have invariably obtained 
the same results on fractionating it. The crude phosphine begins to boil at about 100°, 
the thermometer then rises rapidly to 160°, and from 160°-190° most passes over. The 
residue in the distilling flask decomposes if the distillation is pushed further, and red 
phosphorus separates. All the fractions contain the phosphine, for they all have its 
powerful and characteristic odour, and when mixed with fuming hydriodic acid they 
give its crystalline hydriodate. After repeated rectifications, the lower boiling fractions 
resolve themselves into a liquid, boiling at 110°-1 16°, which is no doubt toluol, while 
the high boiling fractions pass over from 175°-185°, the thermometer being fairly 
constant at 180°-183°. 

We have investigated the high boiling residue (above 190°), and give the results on 
p. 585. 

In view of the difficulty experienced in separating the primary phosphine by simple 
distillation, in a pure state, we decided to obtain the crystallised hydriodate from the 
crude product, and from it the phosphine, but, owing to the bulky nature of that 
compound (i.e., the hydriodate) and its insolubility, we experienced considerable difficulty 
in effecting this. After several experiments, we found that either of the two following 
methods might be employed : — 

( 1 ) The crude phosphine was placed in a retort, and a stream of pure, dry, hydriodic 
acid gas was conducted by a long tube into the body of the retort. As soon as saturation 
appeared to be complete, the retort was heated gradually in an oil-bath to a temperature 
of 160°-180°, a slow current of hydriodic acid passing all the time. The hydriodate then 
sublimed in beautiful colourless scales, and when most had thus volatilised into the neck 
of the retort, the latter was allowed to cool, the hydriodate shaken out, and well washed 
with pure benzol. 

(2) The crude product was mixed with about twenty times its volume of pure dry 
benzol, and the mixture saturated with dry hydriodic acid. It grew warm, and eventually 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 543 

almost solid, from the separated hydriodate. The mass was thrown on to a linen filter and 
thoroughly squeezed, then pressed between filter paper, broken up, and washed with 
benzol, so long as the latter dissolved anything. The benzol was then removed as 
far as possible by pressure between filter paper, and the purified hydriodate dried in 
vacuo. 

The hydriodate prepared by both these methods was snow white, and tolerably per- 
manent in the air ; but if not carefully prepared and thoroughly washed, it rapidly 
became brown. About 60 grms. of the hydriodate thus prepared were placed in a sepa- 
rate funnel, and caustic potash solution added until the funnel was nearly full ; the 
mixture was then well shaken, when the hydriodate rapidly decomposed, and the 
phosphine separated as an oily layer, which floated on the watery liquid containing 
potash and potassium iodide. It was decanted, and submitted to fractional distillation in 
a stream of hydrogen. The thermometer rose rapidly to 178°, then slowly to 190°. It was 
fairly constant from 180°-182° when most distilled; only a little passed from 182°-190°. 
Fraction 178°-190° was redistilled. The thermometer rose at once to 177°, and from that 
temperature to 185° most distilled. The exact boiling point could not be fixed, but it is 
somewhere about 180°-183°. These experiments, conducted with the greatest possible care, 
and repeated two or three times, appear to indicate that the primary phosphine suffers a 
slight decomposition during distillation, but it is also possible that, in spite of the precau- 
tions adopted, the phosphine after all was not absolutely pure. 

Properties. — Monobenzyl phosphine is a colourless, highly refractive liquid, possessing 
a very characteristic and penetrating odour. Its smell remains for days on the hands, 
and in one case it was observed on an instrument months after the latter had 
been handled by one of us, our fingers having been previously in contact with some 
of it. 

Exposed to the air, it at once fumes very powerfully and grows hot. Its vapour indeed 
often inflames spontaneously on leaving a bottle containing it open for some time. The 
product of its oxidation is not a single substance, but contains no less than three different 
bodies. 

Sulphur only acts upon it when the mixture is warmed. Torrents of sulphuretted 
hydrogen are then evolved, and a liquid product is formed. 

It combines readily with hydriodic and hydrobromic acids, and also, though not so 
energetically, with hydrochloric acid. The resulting compounds are crystalline, and are 
very sparingly soluble in a saturated solution of the hydracid. They are volatile, with 
decomposition more or less complete — unless in a stream of gaseous hydracid, when the 
hydriodate and hydrobromate at all events sublime unchanged. They are immediately 
decomposed by water and alkalies. Halogens act violently upon the phosphine, and 
seize upon part (or all) of its hydrogen, the hydracid which is formed then combining with 
the rest of the phosphine to produce its haloid salt. 

Bisulphide of carbon attacks it when the two are heated under pressure, sulphuretted 
hydrogen escapes, and two sulphurised products result. 

VOL. XXXV. PART II. (NO. 15). 4 U 



544 PROF. LETTS AND MR R. F. BLAKE ON 

Chloracetic and bromacctic acids also react with it — the first only when heated with 
the phosphine, the second at ordinary temperatures. 

Chlorocarbonic ether also reacts with it at ordinary temperatures. 

Monobenzyl Phosphine Hydriodate, C 7 H 7 PH 2 HI. — This salt is easily formed, either 
by subliming the phosphine in dry hydriodic acid gas; by saturating a solution of the 
phosphine in benzol with dry hydriodic acid gas; or by dissolving the phosphine in 
warm fuming aqueous hydriodic acid. 

By the first method it is obtained in snow-white scaly crystals like benzoic acid ; by 
the second, as a seemingly amorphous bulky precipitate ; while by the third it is also 
obtained in the crystalline state. 

A specimen prepared by the first method was analysed. 

Analysis. 

0-6090 gave 05630 Agl = -3043015 1 = 4997 per cent. 

Obtained. Calculated for C 7 H 7 PH 2 .HI 

Iodine, . . . 4997 50-39 

The hydriodate, when pure and dry, is permanent in dry air ; but a trace of impurity 
causes it to become brown. It is rapidly decomposed by water, and instantly by caustic 
potash solution. 

Monobenzyl Phosphine Hydrobromate, C 7 H 7 PH 2 HBr. — Hofmann could not obtain 
this compound in the crystalline state, but we found that it could be prepared 
with the greatest ease either by saturating a solution of the phosphine in benzol with 
gaseous hydrobromic acid, or by dissolving the phosphine in the fuming aqueous 
acid. It is also produced when bromine acts upon the phosphine in acetic acid solution, 
and is then precipitated as a colourless crystalline powder. 

Analysis. 

1.0-3639 gave 0-3300 AgBr = 01404 Br = 3858 per cent, 

f 02033 H 2 = 002258 H = 504 
0448 gave j Q . 6799 CQg =0 . 18542 c =41-38 

II. 0-8163 required 38-7 c.c? AgN0 3 =03096 Br =37*90 per cent. 

Obtained. 





I. 


II. 


U1MCU IUL ^7 ix 7 


Bromine, 


38-58 


37-90 


3902 


Carbon, 


4138 




40-97 


Hydrogen, . 


504 


... 


4-88 



I. Obtained by the action of bromine on the phosphine. 
II. ,, ,, hydrobromic acid on the phosphine. 

The salt is insoluble in benzol, and only very slightly soluble in warm fuming hydro- 
bromic acid. Heated in a tube, it sublimes in glittering scales with slight decomposition. 
It is very deliquescent, and decomposes rapidly in contact with water and instantly with 
caustic potash. 

Monobenzyl Phosphine Hydrochlorate, C 7 H 7 PH 2 HC1. — We obtained this salt by similar 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 545 

methods to those which we employed for preparing the two compounds just described. 
On passing gaseous hydrochloric acid into a solution of the phosphine in benzol no effect 
is produced until saturation is complete, then colourless crystalline scales begin to 
form. On shaking the phosphine with a saturated aqueous solution of hydrochloric acid, 
a similar precipitate is produced. We have not analysed the compound, as we did not 
obtain it in sufficient quantity, but its composition cannot be doubted. 

Chloroplatinate. — On mixing the phosphine with an aqueous or alcoholic solution of 
chloride of platinum, a light yellow bulky amorphous precipitate is produced. We did 
not analyse it. 

Action of Air or Oxygen on Monobenzyl Phosphine. — As we have before mentioned, 
the phosphine attracts oxygen with great energy from the air, the temperature rises 
considerably, dense white vapours are produced which occasionally take fire spontaneously. 
The final product of the oxidation is a colourless viscous liquid which refuses to crystallise. 
It dissolves somewhat sparingly in water, and has a strong acid reaction. 

On adding acetate of lead to its aqueous solution a white flocculent precipitate is 
produced which is by no means quite insoluble, so that its bulk diminishes considerably 
on washing. The following results were obtained on submitting it to analysis : — 

Analysis. 

I. 01060 gave 00925 PbS0 4 = 06319 Pb = 5960 per cent. 
II. 04538 „ 0-4035 „ =275 Pb = 608 

I. Prepared from the phosphine obtained from its hydriodate. 
II. Prepared from the crude phosphine. 

Obtained Calculated for 

^ ^p (CVHyHPO^Pb (C 7 H 7 )P0 3 Pb 

Lead, . . 596 60-8 3992 54-97 

It is obvious that the lead salt is neither the benzyl-phosphinite nor benzyl 
phosphinate. On the other hand, the percentage of lead does agree fairly well with 
that required for the formula (C 7 H 7 )PO.Pb — viz., 60 - 0, and this view of its composition, 
namely, that it is a derivative of an oxide C 7 H 7 PH 2 in which the two atoms of hydrogen 
are replaced by one atom of lead, unlikely as it appeared, was supported by the increase 
in weight which a sample of the pure phosphine experienced on spontaneous oxidation — 
a rough experiment giving 14*1 per cent, increase in weight, instead of 12 '9 per cent., the 
calculated amount. 

At the time of our first experiments on this subject we were unacquainted with 
benzyl phosphinic acid, and hence knew nothing about the properties of its lead salt or 
of its other compounds. After it had been obtained, however, and its salts investigated, 
a simple means was at our disposal for ascertaining whether it was present in the product 
of oxidation of the primary phosphine, and for separating it if necessary from other 
substances. For benzyl phosphinate of barium is a highly characteristic salt, being 
much less soluble in boiling water than in cold, and is precipitated almost completely 



f)4() PROF. LETTS AND MR R. F. BLAKE ON 

iii the crystalline state on warming its cold solution, as it is only soluble to the extent 
of less than i per cent, in a boiling solution. 

Accordingly we oxidised a considerable quantity of the primary phosphine — dissolved 
the product in water and neutralised the solution with baryta, when a white flocculent 
precipitate was thrown down, which readily dissolved in acids, and gave a strong 
phosphoric acid reaction with molybdate of ammonia. There can be no doubt that it 
was phosphate of barium. The solution filtered from it was boiled to small volume, 
and gave a crystalline precipitate, which was purified by solution in cold water and 
reprecipitating by boiling. It had the appearance and properties of benzyl phosphinate 
of barium, and was proved by analysis to be that substance. 

Analysis. 

0-6367 lost at 110° C. 00630 H 2 = 989 percent. 

06367 gave 0-4296 BaS0 4 = 0-25259 Ba- 39-67 





Obtained. 


Calculated for C 7 H 7 P0 3 Ba,2H„0 


Water, 


9-89 


1049 


Barium, 


39-67 


39-94 



The concentrated mother-liquors from this salt were syrupy and had a slight odour 

of the primary phosphine. After some time they became granular from the separation of 

a crystalline salt. The semisolid mass was dried for some time in a desiccator, then 

washed with alcohol (in which it was somewhat soluble), and dried in vacuo. A quantity 

of the salt thus purified was dried at 110° and analysed. 

Analysis. 

1098 gave 0-5603 BaS0 4 = 032944 Ba = 3000 per ceut. 

Obtained. Calculated for (C 7 H 7 HP0 2 ) 2 Bu 

Barium, . . . 3000 3064 

In another experiment, the nature of the products of oxidation of the phosphine was 
proved in a similar manner, with this difference, that after separating the phosphate and 
phosphinate of barium, the remaining phosphinite was decomposed by sulphuric acid, 
and the liberated phosphinous acid extracted with ether. The ethereal extract left on 
evaporation a very viscous colourless liquid, which refused to crystallise. It was proved 
to be the phosphinous acid by the production of a number of salts which were analysed, 
and are described on pp. 610-611. 

In a third experiment on the oxidation of the primary phosphine, the product was 
treated in a totally different manner, but with the same results. 

In this experiment the oxidised product was dissolved in water as before, the solution 
boiled and acetate of zinc added. At first a precipitate was thrown down, which redissolved 
as fast as it was formed, but presently on adding more of the zinc salt a permanent 
precipitate was produced, which became pasty on boiling. It was filtered off, 
and the solution concertrated, when a totally different salt was thrown down. 
Analysis proved the first of these salts to be the phosphinate of zinc, the second the 
phosphinite. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 547 

Analysis of No. 1 Salt. 

0-9037 gave 02839 ZnO = 02278 Zn = 25-20 per cent, 

na M o f 01914 H,0 = 0-02126 H= 3-52 „ 

06028 gave j ^^ ^ = ^^ Q = 33 . 66 ^ 

Obtained. Calculated for C 7 H 7 P0 3 Zn,H,0 
Zinc, .... 2520 25"70 

Carbon, . . . 3366 3320 

Hydrogen, . . . 3"52 3-55 

Analysis of No. 2 Salt. 

I. 05145 gave 01133 ZnO = 00901 Zn = 17-66 per cent. 
II. 08757 „ 01946 „ =015616 „ =17-83 „ 

n , 97Q f 02021 H,0 = 0022455 H= 425 „ 

5,173 gave ^^ C 5 2 = -229554 C = 4353 „ 

Obtained. 

, » , Calculated for (C 7 H 7 HPO,).,Zn 

I. II. 

Zinc, . . 1766 17 85 17-33 

Carbon, 4353 44-80 

Hydrogen, 425 4-26 

The experiments described above prove conclusively that when monobenzyl phosphine 
oxidises spontaneously it gives rise to three different products, namely, phosphoric acid, 
benzyl phosphinous acid (C 7 H 7 )HoP0 2 , and benzyl phosphinic acid (C 7 H 7 )H 2 P0 3 . With 
regard to the relative quantities of the three substances, the phosphinous acid is produced 
in by far the largest proportion, and phosphoric acid in the smallest ; the phosphinous 
acid is, in fact, the main product of the reaction. It is probable that phosphinous acids 
are always formed when primary phosphines are spontaneously oxidised. 

H. Kohler and Michaelis* state that phenyl phosphine oxidises almost quantitatively 
in this manner, para-tolyl phosphine also gives rise to the corresponding phosphinous 
acid,t and Moslinger J states that octyl phosphine is probably converted into a 
phosphinous acid by spontaneous oxidation. But, so far as we are aware, no other 
experiments have been made on this subject, which is one of some interest. 

Action of Bromine on the Product of the Oxidation of the Primary Phosphine. — 
In some of our earlier experiments, before the nature of the products of the oxidation of 
the primary phosphine had been ascertained, one of us and W. Wheeler submitted these 
products to the action of bromine. We found that when bromine was added it was 
rapidly decolorised and the mixture grew hot, while hydrobromic acid was evolved, 
and the pungent odour of bromide of benzyl became manifest. Excess of bromine 
was added, and the mixture allowed to remain undisturbed for some time. A crop of 
crystals then separated, which dissolved both in water and ether, and were obtained 
colourless by recrystallisation. These crystals on analysis gave the following results : — 

* Kohler and Michaelis, Berichte, x. (1877) p. 810. 

t Michaelis and Panek, Annalen, 212, p. 234. J Moslinger, Berichte, ix. (1876) p. 1008. 






T)48 PROF. LETTS AND MR R. F. BLAKE ON 

Analysis. 

0221 gave 0166 AgBr = 07065 Br = 3L9 per cent. 

noszL f 06245 C0 o =0-18735 C =489 

U ZV* gave ( 01535 u ^ =0 . 01705 H = 8 . 9 () 

Obtained. Calculated for (C 7 H 7 ) 3 P 2 Br 2 H 3 
Carbon, . . . 489 490 

Hydrogen, . . . 4 - 9 4'7 

Bromine, . . . 319 311 

In another experiment, conducted in a similar manner, we also obtained a solid 

crystalline substance, which, after crystallisation from water, consisted of scales with 

mother-o'-pearl lustre, and a melting point of 176° C. A bromine determination gave the 

following result : — 

02423 gave 0176 AgBr = 0849 Br = 309 per cent. 

In a third experiment completely different results were obtained. In this experiment 
a quantity of the primary phosphine, which had been kept for some years in a loosely- 
corked flask, was treated with bromine, when, as in previous experiments, abundance of 
hydrobromic acid was evolved, and the mixture grew very hot. The product was heated 
for a considerable time on a water-bath to get rid of hydrobromic acid. The residue 
had an intolerable odour of bromide of benzyl. On treatment with water some of it 
dissolved, but an oily liquid remained, which solidified on cooling. The aqueous solution 
gave the reactions of benzyl phosphinic acid. The oily liquid had acid properties. It 
was dissolved in baryta water, the solution filtered, and precipitated by hydrochloric acid. 
The resulting crystalline precipitate was recrystallised from hot alcohol, in which it readily 
dissolved. Its corrected melting point was now found to be 176° C. Its analysis gave 
the following results : — 



Analysis. 



( 03920 C0 2 = 0T06909 C = 6397 per cent. 
■] 0-1014 H 2 6 =0011266 H= 674 
1 00732 Mg,P 2 7 = 0-020443 P = 1223 



Obtained. Calculated for C^li^PjOj 
Carbon, .... 63-97 6393 

Hydrogen, . . . 674 6"96 

Phosphorus, . . . 12-23 1270 

A quantity was converted into its barium salt, and the latter analysed. 

Analysis. 

0-3214 i lost at l10 ' °' 0328 = 10 " 20 P er cent - H 2° 

t gave OT072 BaSO 4 = 0063031 Ba = 19-61 per cent. Ba 

Obtained. Calculated for C 6 H 32 P 2 6 Ba,4H 2 

Barium, .... 1961 1971 

Water, .... 1020 1036 

As we found that benzyl phosphinic acid is not attacked by bromine at ordinary 
temperatures, the substances which we obtained by the action of bromine on the oxidised 
phosphine are probably formed from benzyl phosphinous acid. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 549 

The experiments just described shows that the reaction is of a highly complex nature, 
and that it varies with the conditions of the experiment. 

As the substances resulting from the reaction were obtained in very small quantity, 
barely sufficient for their analysis, we were unable to study their properties, and must 
remain in doubt regarding their exact nature. 

Action of Sulphur on Monobenzyl Phosphine. — A preliminary experiment showed 
that no action occurs in the cold, but on warming the two bodies together the sulphur 
dissolves and abundance of sulphuretted hydrogen escapes. 

About 4 grms. of the phosphine were placed in a test-tube full of carbonic anhydride, 
and heated by immersing the tube in a water-bath. Powdered sulphur was then added, 
and sulphuretted hydrogen was evolved in abundance — to such an extent, indeed, that 
at one time the liquid frothed over into the water-bath. Excess of sulphur was added to 
the liquid, which had thus frothed over (B), and the same was done with the liquid in 
the tube (A) — a very considerable quantity being required in both cases. 

As soon as the evolution of sulphuretted hydrogen ceased, the liquid was allowed to 
cool. It was colourless and viscous. It was boiled with water, when sulphuretted 
hydrogen was again evolved — nearly the whole (of the viscous liquid) dissolved, forming 
a strongly acid solution. This was neutralised with baryta, and the solution heated, 
when a crystalline salt separated very much like benzyl phosphinate of barium, and was 
at first considered to be that substance. 

The product B gave a similar salt. The two salts were purified by dissolving them in 
cold water and then heating the solution when they were deposited in the crystalline state. 

Analysis of Barium Salt from A. 

0-4710 gave 03184 BaS0 4 = 0187213 Ba = 3975 per cent. 
04710 lost at 110° C. 00253 H 2 = 537 

Analysis of Barium Salt from B. 

0-5691 gave 0-3878 BaS0 4 = 0228019 Ba =4006 per cent. 
0-5691 lost at 110° C. 00289 H,0 = 507 ;, 





Obtained. 


Calculated for (C 7 H 7 )P 




A. B. 


Barium, 


39-75 4006 


4017 


Water, 


5-37 507 


527 



On attempting to obtain the acid by decomposing the barium salt with sulphuric 

acid, sulphuretted hydrogen was evolved — no doubt, with formation of benzyl phosphinic 

acid — 

C 7 H 7 P0 2 SH 2 + H 2 = C 7 H 7 P0 3 H 2 + H 2 S . 

The results of this experiment indicate that when the phosphine is warmed with 
sulphur, pyro-benzyl-thio-phosphinic acid is formed, 

2C 7 H 7 PH 2 + 5S = (C 7 H 7 ) 2 P 2 S 5 H 2 + H 2 S , 



550 PROF. LETTS AND MR R. F. BLAKE ON 

and this with water is decomposed into mono-thio-phosphinic acid, with evolution of 
-ulphuretted hydrogen, 

(C 7 H 7 ) 2 P 2 S 5 H 2 + 6H 2 = 2C 7 H 7 PS(OH) 2 + 5H 2 S . 

In a second experiment we attempted to obtain derivatives of the pyro acid. A 
quantity of the phosphine was heated as before with excess of sulphur until sulphuretted 
hydrogen ceased to be evolved. The product deposited a minute quantity of colourless 
plates, which were not obtained in sufficient quantity for analysis. 

A quantity of the product was boiled with a solution prepared by saturating barium 
hydrate with sulphuretted hydrogen, and then adding an equal quantity of the hydrate. 
Sulphuretted hydrogen was evolved, and when the solution had become very concen- 
trated, it deposited colourless crystals on cooling, which were collected and analysed. 



Analysis. 



0T516 i lost at 110 ° C> °' 0096 = 6 * 33 P er cent water - 

j gave 00647 BaS0 4 =0 038042 Ba = 2509 per cent, barium. 





Obtained. 


Calculated for (C 7 H 7 ) 3 P 2 S 3 OoBa,2H 2 


Barium, . 


25-09 


2513 


Water, 


6-33 


6-60 



The production of the pyro acid is thus to a certain extent confirmed, its barium 
salt, which is no doubt formed in the first instance, being subsequently decomposed by 
water, thus — 

(1) (C 7 H 7 ) 2 P 2 S 5 H,+BaS = (C 7 H 7 ) 2 P 2 S 5 Ba +H 2 S 

(2) (C 7 H 7 ) 2 P 2 S 5 Ba+ 2H 2 = (C 7 H 7 ) 2 P 2 S 3 2 Ba+ 2H 2 S . 

We may also mention, that on dissolving the product of the action of sulphur on the 
phosphine in alcohol and adding mercuric chloride, a white amorphous precipitate was 
produced in abundance. We did not analyse it, as it darkened slowly on drying. 

We have already mentioned that Michaelis studied the action of sulphur on phenyl 
phosphine and obtained two products — a liquid (C 6 H 5 )PH 2 S, and a solid (C 6 H 5 P) 3 S. The 
experiments we have just described show that benzyl phosphine behaves in a totally 
different manner, and they also prove that sulphur and oxygen do not react in an 
analogous manner upon a primary phosphine. 

Action of Halogens on Monobenzyl Phosphine. — If bromine is mixed with the 
phosphine an explosive action occurs, and torrents of hydrobromic acid are disengaged. 
But when the vapour of bromine comes in contact with the base, the action proceeds 
quietly, and a white solid is produced. If both the bromine and phosphine are dissolved 
in glacial acetic acid, the reaction is completely under control. If the vessel in which 
the mixture is made is kept cool and the bromine gradually added, a colourless crystalline 
salt is produced, and its amount increases as the bromine is added, up to a certain point, 
but it then diminishes. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 551 

A quantity of this solid was filtered off, washed with glacial acetic acid, and then 
boiled with acetic acid until it dissolved. In dissolving it effervesced from the escape of 
hydrobromic acid. As the solution cooled colourless plate-shaped crystals separated. 
These were dried in vacuo, and analysed. 



Analysis. 



0-3639 gave 03300 AgBr = 01404 Br = 3858 per cent. 

80 j gave 0-2033 H 2 = 002258 H - 5'04 per cent. 
( „ 0-6799 C6„= 0-18542 C = 41-38 „ 





Obtained. 


Calculated for (C 7 H r )PH 2 .HBr 


Bromine, 


38-58 


3902 


Carbon, 


4138 


40-97 


Hydrogen, 


5-04 


4-88 



The crystalline solid was thus shown to be the hydrobromate of the phosphine, and 

all its properties pointed to the same conclusion. One or both of the following reactions 

must have occurred : — 

(C 7 H 7 )PH 2 +Br 2 =HBr + (C 7 H 7 )PHBr 

(C 7 H 7 )PH 2 + 2Br 2 = 2HBr + (C 7 H 7 )PBr 2 . 

The hydrobromic acid set free combining with the phosphine remaining in excess, and 
thus the quantity of hydrobromate produced would increase until the bromine began to 
be in excess, when, no doubt, it (the hydrobromate) was also attacked in the same way as 
the phosphine itself. 

The mother-liquors from which the hydrobromate had separated ought to have 
contained one or other of the two brominatecl derivatives whose formulae we have 
written above, but we had not sufficient of the product to be able to isolate either of 
them. 

On evaporation they left a solid crystalline mass, which dissolved in water (leaving a 
few oily drops), and when the solution was neutralised with baryta and boiled, a crystal- 
line salt was precipitated, having the appearance of benzyl phosphinate, but we did not 
obtain it in sufficient quantity for analysis. 

Action of Bisulphide of Carbon on Monobenzyl Phosphine. — The following experi- 
ments were made : — 

(1) 2 grms. of the pure phosphine were heated in a sealed tube with 2 grms. of 
bisulphide of carbon at 120° C. for two days. The contents of the tube then consisted 
of a viscous colourless substance and a number of colourless needle-shaped crystals. On 
opening the tube a considerable quantity of sulphuretted hydrogen escaped. 

(2) About 4 "5 grms. of the phosphine and about 9 grms. of the bisulphide of carbon 
were heated for two days at 130°-160°, when exactly the same phenomena were observed. 
The contents of each of the two tubes were separately treated with bisulphide of carbon, 
which dissolved the viscous substance, but left the crystals. The latter were repeatedly 
washed with bisulphide of carbon, then dried and analysed. 

VOL. XXXV. PART II. (NO. 15). 4 X 



55*2 PROF. LETTS AND MR R. F. BLAKE ON 

Analysis of Crystalline Product from Experiment 1. 

~ . , , , AS)1>7 - f 0-107 H.,O = 00118 H = 542 per cent. 

Carbon and hydrogen, 0-2175 gave { ^^ ^ = ^^ Q = ^ 

Analysis of Crystalline Product from Experiment 2. 

n , , , , aoqka f 01336 H„0 = 0014844 H = 520 per cent. 

Carbon and hydrogen, 0-2850 gave j ^^ ^^^ c= ^. 08 

0bt f ned - Calculated for 

' j Yl * (C 7 H r )PH 2 S (C 7 H 7 )PHS C r H r PS 

Carbon, . . 5430 54'08 5385 54-19 54-54 

Hydrogen, . . 5-42 520 5"76 516 4-54 

The bisulphide of carbon washings from the crystals were warmed to get rid of the 
bisulphide, and left a pale yellow gummy mass, which was insoluble in water and ether, 
and very sparingly soluble in alcohol. No definite products could be obtained from this 
gummy mass by the action of various reagents. It was, however, found that it was 
soluble to a certain extent in boiling glacial acetic acid, so the whole of the product from 
the first experiment was boiled with a large quantity of the acid. The filtered solution 
deposited oily droplets on cooling, which eventually formed a viscous mass exactly like 
the original substance (specimen A). The acetic solution filtered from this was 
evaporated to dryness in a water-bath, and heated until the whole of the acetic acid 
had volatilised. A gummy mass remained also like the original product (specimen B). 

The whole of the product from the second experiment was repeatedly boiled with 
alcohol, then washed with cold alcohol and dried (specimen C). 

The three products were then analysed. 

Analysis. 

A. 



ft . Q „ f 01982 H = 002202 H = 505 per cent. 

4355 gave| . 8280 ^ =0 . 22 582 C = 51-80 „ 

03463 ?ave I 01595 H 2 = 001772 H= 510 per cent. 
3463 gave j ^^ Qa = ^^ c = ^ 

C. 

f\ mak $ °' 1287 H 2° = 00143 H = 540 per cent. 

.045 gave j Q ^ 8Q ^ = 01333g c = 5Q ^ 



4889 C0 2 =013333 C =5040 


j? 


Obtained. 

A. B. C. 

51-80 5150 5041 

505 510 540 


Calculated for 
(C 7 H 7 PHCS) 2 S 

5245 
4-37 



Carbon, 
Hydrogen, . 

To the best of our belief, the action of bisulphide of carbon on a primary phosphine 
has been studied in only one other case, viz. , in that of phenyl phosphine. 

Michaelis and Dittler found that the two substances react under pressure at 150° 
with evolution of sulphuretted hydrogen, and formation of a resin having the composition 
(C 6 II n PHCS) 2 S, the reaction occurring according to the equation, 

2C G H 5 PH 2 + 2CS, = (C 6 H.PHCS) 2 S + H 2 S . 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 553 

The compound which they name phenyl-diphospho-sulpho-carbamic acid dissolves in 
alkalies, and is reprecipitated unchanged on the addition of an acid. Its discoverers do 
not mention any other substance as being formed at the same time analogous to the 
crystalline body which we obtained. 

We cannot be at all certain of the composition of either of the two substances ob- 
tained by us. As regards the crystalline product, the analytical results agree fairly 
well with the formula (C 7 H 7 ) 2 P 2 H 2 S2, but we did not obtain it in our experiments on 
the direct action of sulphur on the phosphine, as we should have done in all probability 
if the above formula were correct. It was produced in too small a quantity for further 
investigation. With respect to the viscous product, we cannot believe that it is 
analogous to the phenylated body, for it has no acid properties, and the analytical 
results are not sufficiently in accordance with the formula. For the present we remain 
in doubt as to the nature of the substances. 

Action of Sulphuric Acid on Monobenzyl Phosphine. — If sulphuric acid (con- 
centrated) is added to the phosphine cautiously, and the mixture kept cold, it solidifies 
to a crystalline mass ; but if it is not cooled a violent action soon occurs, and sulphurous 
anhydride is evolved in abundance. It may be inferred that the phosphine forms a 
crystalline sulphate which is very unstable, and probably decomposes into benzyl 
phosphinic acid and sulphurous anhydride. 

Attempts to obtain the sulphate for analysis were unsuccessful. 

Action of Chloracetic Acid on Monobenzyl Phosphine. — Three separate experiments 
were tried, with similar results in each case. When equivalent quantities of the phos- 
phine and chloracetic acid are mixed no action occurs, but some of the acid dissolves. 
On gently warming the mixture the whole of the acid dissolves, but separates out again 
on cooling. If the two substances are heated in a sealed tube at 120° for some hours, a 
dark brown viscous mass results, and on opening the tube pressure is noticed, due to 
carbonic anhydride. The product undoubtedly contains chloride of acetyl, for it has its 
very characteristic odour, and fumes in moist air. On removing it from the product by 
shaking the latter with ether, and then treating the residue with water, a strongly acid 
solution results, while a little tarry matter remains undissolved. On neutralising the 
acid solution with baryta and then warming, a crystalline salt is precipitated. This we 
submitted to analysis (after recrystallisation), and found it to consist of benzyl phosphina.te 
of barium. 

Analysis. 

. 4794 ( lost at 110° C. 0047 H 2 = 9"80 per cent. 

t gave 03189 BaS0 4 = 0-1875077 Ba = 3911 

Obtained. Calculated for (C 7 H 7 )P0 3 Ba,2H 2 

Barium, .... 3911 3994 

Water, .... 980 1049 

We think that the main course of the reaction must proceed as follows : — 

C 7 H 7 PH 2 +3CH 2 C1.C0 2 H = C^PHA+SCH .COC1 . 



o.")4 PROF. LETTS AND MR R. F. BLAKE ON 

It is remarkable that the phosphine should thus seize upon the oxygen of the 
c;irboxyl group, and not upon the halogen. 

It is probable that another reaction occurs to a slight extent, viz.: — 

C 7 H 7 PH 2 + CH.,C1 - COOH = (C 7 H 7 )(CH,)PH,HC1 + CO , . 

This equation, at all events, accounts for the liberation of carbonic anhydride, and is 
supported to a certain extent by the experiments which one of us has made on the action 
of chloracetic acid on triethyl-phosphine, or rather upon that of heat upon the product.* 

Action of Bromacetic Acid upon Monobenzijl Phosphine. — We have not fully 
investigated the nature of this action, but we found that when the two substances 
were mixed, the bromacetic acid dissolved, and in a short time a. crystalline product 
resulted. In about half an hour's time, however, an explosive decomposition of this 
substance occurred spontaneously, torrents of hydrobromic acid coming off, and a viscous 
liquid resulting, which was insoluble in water. We did not investigate the product, as 
nearly all of it was shot out of the vessel in which the experiment was made when the 
explosive action occurred. It is probable, we think, that a, product of addition is first 
formed, which may then decompose in the following way : — 

7 H%P<^2-COOH = C , 7 H 7 \ >p _ CHCOOH + HBr 

Action of Chlorocarhonic Ether on Monobenzijl Phosphine. — When the two sub- 
stances are sealed up together a reaction gradually occurs, and thin quadratic plates of 
fair size are slowly deposited. On opening the tube pressure is observed. We could 
not obtain the crystals in a sufficiently pure state for analysis. They were so deliquescent 
that on attempting to dry them on filter-paper, they almost instantly deliquesced. They 
dissolved in alcohol, ether, and benzol. Alcoholic chloride of platinum gave a yellow 
slimy precipitate ; alcoholic corrosive sublimate, a, white amorphous precipitate, which 
soon turned grey. On warming them with baryta solution, carbonate of barium was 
precipitated. We think that these properties indicate that the substance is a product of 

addition — 

C 7 H 7 . /C00C 9 H-, 

H / X C1 . 

Hofm ann's " Dibenzyl " Phosphine. 

In his well-known researches on the phosphines, Hofmann has apparently shown 
that when an alkyl haloid is heated with phosphonium iodide and oxide of zinc, 
primary and secondary phosphines alone result ; whereas when an alcohol is heated 
with iodide of phosphonium alone, tertiary and quaternary phosphines are formed 
exclusively. Thus the two reactions arc complementary to each other. Among the 

* Letts, these Transactions, xxx. part 1, p. 285. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 555 

series to which he extended his investigations was that of benzyl, and in the paper 
published in the Berichte, to which we have already referred, he describes the 
preparation and properties of dibenzyl phosphine. He isolated it from the product of 
the sealed tube reaction (occurring between benzyl chloride, phosphonium iodide, and 
oxide of zinc), but it is not quite clear from the original memoir whether he obtained it 
from the residue left on distilling the primary phosphine in a current of steam from the 
product of the reaction, or the residue left on distilling the crude primary phosphine 
itself. He states that a crystalline substance remains after distilling off the primary 
phosphine, which solidifies, especially (zumal) in contact with solid potash, and that after 
pressing it between linen, dissolving in alcohol, and decolorising the solution with 
charcoal, crystals are obtained, which after a single recrystallisation from alcohol are 
perfectly pure. 

The crystals melted at 205°, and their analysis gave the following numbers : — 

Obtained. 





1. 


II. 


III. 


,o '"' ii,ku "" ^ 


Carbon, . 


7S-75 


78-37 




78-50 


Hydrogen, 


6-99 


677 




701 


Phosphorus, . 






136 


1449 



He says : — " With the entrance of the second group the basic properties, which in 
benzyl phosphine appear in a perfectly definite manner, have entirely disappeared. 
Dibenzyl phosphine dissolves in no acid, neither have I succeeded in obtaining its 
platinum salt. The aromatic secondary phosphine thus distinguishes itself very materially 
from the secondary phosphines of the ethyl and methyl series." 

In a paper read before this Society, one of us, in conjunction with W. Wheeler, 
described further investigations on this body, and we showed that it forms a series of 
compounds of a somewhat remarkable nature for a secondary phosphine. We isolated it 
as follows : — After submitting the contents of the sealed tubes to the action of steam 
and water (to drive off the primary phosphine), the residue, consisting of a brownish 
viscous mass, was boiled with aqueous potash for some time, and the viscous insoluble 
mass which remained was then extracted with hot alcohol. The solution on cooling- 
deposited colourless needle-shaped crystals strongly impregnated with oily matter, and 
these by repeated recrystallisation from spirit were obtained pure. 

The remarkable nature of this substance, on the assumption that it is in reality a 
secondary phosphine, coupled with some other facts which attracted our attention, 
gradually led us to suspect that it was not dibenzyl phosphine at all, but the oxide of 
tribenzyl 'phosphine. Accordingly, the investigation was reopened with Mr K. F. Blake, 
and its course has been as follows : — 

1. On carefully recrystallising the substance from alcohol, its corrected melting- 
point was found to be 215°-215°*5 C, while that of two specimens of oxide of tribenzyl 
phosphine (prepared by two different methods) was found to be the same. 



f>56 PROF. LETTS AND MR R. F. BLAKE ON 



2. Very little difference exists in the percentage amount of carbon and hydrogen in 
dibenzyl phosphine and oxide of tribenzyl phosphine, as the following numbers show :• — ■ 





(C 7 H 7 ) 2 H1> 


(C 7 H 7 ) s PO 


Carbon, 


78-50 


78-75 


Hydrogen. . 


701 


6-56 



Consequently it would not be possible to decide with absolute precision between the 
two substances by a mere combustion. On the other hand, there is a considerable 
difference between the two bodies in their percentage of phosphorus — 

(C ; H 7 ) 2 HP (C 7 H 7 ) 3 PO 

Phosphorus, . . . 1448 969 

Unfortunately, however, as we have again and again found, the processes for phosphorus 
determinations in ordinary organic substances are absolutely untrustworthy when 
applied to phosphines. A new method was therefore necessary, and after many trials 
we believe we have found one which is perfectly accurate, trustworthy, and capable of 
general application. It is extremely simple, though somewhat tedious in carrying out. 
It consists in making an ordinary combustion of the substance with pure oxide of 
copper, and afterwards dissolving the contents of the combustion tube in nitric acid, 
and determining the phosphorus with molybdate of ammonia, &c. Applying this 
method to the analysis of the supposed dibenzyl phosphine, we obtained the following 
results (IV. and V.). 

We give at the same time the determinations of phosphorus made both by Hofmann 
(1.) (by a method not described) and by one of us and W. Wheeler (II. and III.), by 
burning the substance with lime in a stream of oxygen, dissolving and titrating with 
.standard uranium solution, 

Obtained. 



I. II. III. IV. V. . 

Phosphorus, . . 136 1435 1500 986 998 

As we have before stated, many of the compounds of the supposed dibenzyl phosphine 
were prepared and analysed, and although their composition appeared remarkable for 
the compounds of a secondary phosphine, the analytical results agreed fairly well in 
most cases with formulae of derivatives of dibenzyl phosphine. We even concluded, 
from the existence of these compounds, that the exceptional properties of the phosphine 
were due to its consisting of a double molecule, viz.: — 

(C 7 H 7 ) 2 H S P = P^C 7 H 7 ) 2 H. 

It is a remarkable coincidence that the analytical results agree equally well for com- 
pounds of the oxide of tribenzyl phosphine. This fact has added considerably to the 
difficulties and uncertainty of the research, but at the same time it has caused us to 
exercise great caution in the inferences to be drawn from the analyses. 

Subjoined is a brief account of the chief compounds we have examined of Hofmann's 
dibenzyl phosphine, and we give at the same time the results obtained (in some cases) 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 557 

with compounds of what was known to be oxide of tribenzyl phosphine. We may 
add that many of the compounds are unstable, and their composition frequently varies 
with the method employed in their preparation. In many cases indeed it appears to be 
almost impossible to obtain them pure. 

Bromide. — This compound, which is highly characteristic, was obtained by adding 
bromine to a solution of the body in glacial acetic acid. It usually crystallises in 
yellow needles, which under the microscope appear to consist of aggregations of minute 
rhombohedra. It is unstable, and gives off bromine when boiled with water or glacial 
acetic acid, and its composition varies considerably (see p. 625). 







Prepared from 

A 




Bromine, . 
Carbon, 
Hydrogen, . 


Hofmann's Dibenzyl Phospli 

Obtained. 
26-7* 
565 + 
5-3 J 


:ine. Oxide oi 
Calculated for 

A 


: Tribenzyl Phosphine. 

Obtained. 
28-4 f 
569 
4-9 


Bromine, 
Carbon, 
Hydrogen, . 


{(C 7 H 7 ) 2 HP} 2 ,Br 2 
27-2 
571 
51 


7(C 7 H r ) 3 PO,5Br„ 
2631 
5800 

4-88 


5(C 7 H 7 ) 3 PO,4Br, 
28-5 
56-3 

4-7 



Iodide. — Prepared like the bromide. It crystallises in minute red crystals of the 
same colour as ferricyanide of potassium. 

Calculated for 

Obtained. , * 

{(C 7 H 7 ) 2 HPM, 7(C 7 H 7 ) 3 PO,5L, 5(C 7 H 7 ) 3 PO,4l 2 
Iodine, . . 3686 3724 36-18 38*84 

Chloride. — Obtained by passing chlorine into a solution of the body dissolved to 
saturation in warm acetic acid. It crystallises when the solution cools in pale yellow 
crystals, much like pentachloride of phosphorus in appearance. The compound is most 
unstable, and loses chlorine rapidly in vacuo, and probably also when air-dried. 

Calculated for 

Obtained. , A * 

{(C 7 H 7 ),HP} 0) C1 2 7(C 7 H 7 ) 3 PO,5Cl 2 

Chlorine, . . 1200 1423 1468 

Hydriodate. — Obtained by saturating a solution of the body in glacial acetic acid 
with hydriodic acid gas, and separated as the solution cooled in colourless crystals. 

Calculated for 
Obtained. 



2(C 7 H 7 ) 2 PH,HI 3(C 7 H 7 ) 3 PO,2HI 
Iodine, . . {21-5} 22 ' 8 21 ' 05 

* Mean of ten determinations of different preparations of the compound. The extremes were 25 - 9 to 27'2. 

t An old analysis by one of us and N. Collie, these Transactions, xxx. part 1. + Mean of three determinations. 



558 PROF. LETTS AND MR R. F. BLAKE ON 

Hydrobromate. — Obtained in the same way as the hydriodate : — 

< Calculated for 
Obtained. 



2(O.H.) 2 HP,HBr 3(0-1 l 7 ),PO,2HBr 

{: 



(1) 195) 

Bromine,. . -J (2) 205 J- I5i) 1444 



(3) 163) 

Platinum Salt. — Prepared by mixing alcoholic solutions of chloride of platinum and 
of the substance. The compound crystallises out in minute leaflets. Its composition 
varies with the conditions under which it is prepared. 

Obtained from 





Hofmann's 
I. 


Dibenzyl Phosphine. 
II. III. 


Oxide of Tribenzyl Phosphine. 
IV. V. 


Carbon, 


59-5 


565 


58-5 


58-9 


59-4 


Hydrogen, . 
Chlorine, 
Platinum, . 


5-8 
128 


53 
131 


5-9 
131 


53 
( 12-4 
\ 126 


54 



(I., II., III., IV., and V. were all different preparations.) 









Calculated for 

* 




5(C ; 


H 7 ) 2 PH,PtCl 4 




4(C T H 7 ) 3 PO,2HCl,PtCl 4 


Carbon, 


, 


595 




595 


Hydrogen, . 


. 


53 




50 


Chlorine, 








11-7 


Platinum, 




140 




12-5 



Nitro- Compound. — Obtained by dissolving the body in cold fuming nitric acid, and 
then precipitating with water. Amorphous. 

Obtained. Calculated for 

(With Hofmann's , * , 

Dibenzyl Phosphine.) (C 7 H 6 N0 2 ) 2 HP (C 7 H NO 2 ) 3 PO 

Carbon, . . . 5523 55-26 5538 

Hydrogen, . . 430 4'27 395 

Double Salt with Iodide of Zinc. — This compound separates out when fairly strong 
alcoholic solutions of the body and iodide of zinc are mixed, in tufts of characteristic 
needles. 

Obtained from 



Hofmann's Dibenzyl Oxide of Tribenzyl Phosphine. 

Phosphine. I. II. 

Iodine, . . . 26\57 260 259 

Calculated for 



3(C r H 7 ) 2 HP,ZrJ, 2(C 7 H 7 ) 3 PO,ZnI 2 

Iodine, . . . 26 : 43 2648 

Action <>/ fused Potash on Hofmann's " Dibenzyl Phos2)hiue." — In the paper by 
one of us and W. Wheeler, already referred to, the statement is made that when 
Hofmann's dibenzyl phosphine is heated with caustic, potash, or soda, " it fuses and 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 559 

floats on the surface of the melted alkali. No violent action occurs, but on cooling the 
mixture and treating it with water the greater portion dissolves, and acids then precipi- 
tate a flocky crystalline substance, which is dibenzyl phosphinic acid." In corroboration 
of this statement, the melting point of the acid and analyses of its lead and barium salts 
were given, all in accordance with the required numbers. After we had satisfied 
ourselves that Hofmann's dibenzyl phosphine was oxide of tribenzyl phosphine, and 
nothing else, this reaction recurred to our minds as a further and very striking excuse 
for the mistake which we (and Hofmann) had fallen into, and we thought it of im- 
portance to verify the previous observation. This we have accordingly done, both with 
Hofmann's dibenzyl phosphine and with a specimen of oxide of tribenzyl phosphine 
prepared by a different process. 

The phenomena observed were exactly the same as those previously described, and 
the melting point of the acid obtained after two recrystallisations from alcohol was found 
to be 192° C. (corr.), which is the melting point of pure dibenzyl phosphinic acid. Our 
previous observations are thus fully confirmed. 

The reaction in all probability occurs as follows : — 

(C 7 H 7 ) 3 PO + KHO = C 7 H 8 + (C 7 H 7 ) 2 KP0 2 . 

Investigation of the Product of the Sealed Tube Eeaction for the Secondary 

and Tertiary Phosphine. 

The occurrence of oxide of tribenzyl phosphine among the products of the sealed 
tube reaction led us to suspect that the tertiary phosphine had been formed in the 
first instance, but was subsequently oxidised by the air, as in isolating the oxide all 
the operations had been performed in open vessels. Quite accidentally, our suspicions 
received strong confirmation. 

In one of our later experiments, after the primary phosphine had been distilled 
from the crude product of the reaction in a current of steam, the viscous mass remain- 
ing was treated with an alcoholic solution of potash, instead of an aqueous solution as 
we had used in our earlier experiments. 

After boiling for about four hours with alcoholic potash, in a flask fitted with an 
upright condenser, the solution was filtered from oxide of zinc, &c, and on cooling 
deposited feathery crystals like sal-ammoniac, and not at all like the needle-shaped 
crystals of oxide of tribenzyl phosphine. The liquid was decanted from these crystals, 
and to remove adhering potash they were washed with cold water, and were then 
pressed between filter-paper to dry them. On unfolding the paper, the crystalline mass 
grew very hot, and after repeated recrystallisation from alcohol (in an open vessel), the 
resulting crystals were needle-shaped, and had a melting point of 21 5° '5 (corr.). 

The conclusions we drew from this result were that the tertiary phosphine is formed 
in the reaction, and that it is a solid crystalline body which rapidly absorbs oxygen from 
the air with disengagement of heat, and is eventually converted into its oxide. 

VOL. XXXV. PART II. (NO. 15). 4 Y 



560 PROF. LETTS AND MR R. F. BLAKE ON 

Another experiment was made in a similar manner, when the same phenomena 
recurred, the crystalline cake growing so hot that we thought it would ignite. 

The problem we had now to solve was to isolate the tertiary phosphine from the 
numerous bye products, and this we foresaw would be no easy task, as all the operations 
would have to be performed in closed vessels and in an inactive gas. We scarcely 
hoped to be able to purify the phosphine by distillation, as it would almost certainly 
have a boiling point higher than that of triphenyl phosphine, which distils above 360°. 
The only available plan for its purification appeared to be crystallisation from alcohol. 
Accordingly, we constructed an apparatus for liberating the phosphine, and crystallising 
it from alcohol in hydrogen. (Fig. 2 of the Plate.) 

Experiment 1. — Twelve sealed tubes, each charged with 16 grms. phosphonium iodide, 
12 grms. benzyl chloride, and 4 grms. zinc oxide, were heated for six hours at 120°. Their 
contents were then transferred, as we have described on p. 541, to a flask, and the primary 
phosphine liberated by water and steam and distilled off. The residual brown resinous 
mass was steamed three times with water, and was then transferred to the flask A, and 
drained from water as far as possible. 50 grms. of potash were dissolved in 400-500 c.c. 
of alcohol ; the flask A connected with the condenser, filled with hydrogen, and the 
potash solution added through C. D was then closed (with an india-rubber tube and 
a glass rod), and C left open for the escape of hydrogen. The mixture in A was then 
boiled for about three hours, during which a gentle stream of hydrogen was allowed to 
flow through the apparatus. C was now closed. D connected with E, and pushed down 
so that its end was at the bottom of A. The hydrogen thus forced the liquid into F 
(previously filled with the gas), when it was filtered into H. As it cooled crystals 
separated, but in addition an oily liquid also. After the lapse of some time H was closed 
by the pinchcock G, F disconnected, the ends of K and I closed, and H inverted. The 
alcoholic mother-liquors were separated from the crystals and oily liquid by attaching K to 
the hydrogen generator, and opening G, when they were forced out. The crystals and oily 
liquid were washed twice with water added through G. Next, after draining off the 
water, alcohol was run in and the crystals boiled with it, while a stream of hydrogen passed. 
On cooling both the crystals and oily liquid separated out again. The mother-liquors 
were drained off as before, and the crystals and oil dried in a stream of hydrogen (H 
being placed in boiling water)., 

An attempt was then made to separate the phosphine by distillation in vacuo, but 
without success. A little liquid passed over at first, but free phosphorus soon separated, 
and the temperature suddenly rose above 360°, with complete decomposition. 

There had remained in the funnel F, in this experiment, a considerable quantity of 
a dirty oily liquid mixed with some crystals. It was washed two or three times with 
water, and was then examined. 

(1) A portion mixed with iodide of methyl soon reacted, the mixture grew very hot 
and entered into ebullition, and after a short time an oily liquid resulted. 

(2) Another quantity was mixed with crystallised iodide of benzyl, when a good deal 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 561 

of heat was also developed, sufficient indeed to cause the water mechanically mixed 
with the product to boil. On cooling the product partly solidified. By recrystallising 
it from alcohol several times, colourless rhombohedral crystals were obtained, identical 
in form with tetrabenzyl phosphonium iodide, and having a melting point of 18 9° '5 
(corr.) (that of the pure iodide being 189°). 

These results pointed to the conclusion that after all tribenzyl phosphine is not a 
solid, but an oily liquid, which is sparingly soluble in alcohol, and that the spontaneous 
oxidation and heating of the crystalline cake which we have mentioned was due to the 
oxidation of liquid tribenzyl phosphine adhering to the solid oxide. In our next experi- 
ment we proceeded somewhat differently. 

Experiment 2. — The contents of twelve sealed tubes were steamed, &c, and 30 grms. 
of crude primary phosphine obtained. The viscous residue was twice steamed with fresh 
quantities of water to remove the hydracids and zinc salts. It was then boiled in the 
flask A with 50 grms. of caustic potash and about 300 c.c. of alcohol for an hour. The 
tube D was now pushed down below the alcbolic solution, while L was connected with 
a condenser and the condenser B removed, and a cork substituted for it. Steam was next 
blown through D until all the alcohol was distilled off and a good deal of water had 
condensed in A. At the bottom of this was a layer of oxide of zinc, while above it an 
oily layer floated. The whole was allowed to cool in a current of hydrogen, when the 
oily layer solidified. A was now inverted, and a current of cold water run in through D, 
and allowed to run out through L. By this means the solidified oily matter was 
thoroughly washed without coming in contact with air, and freed from potash and oxide 
of zinc. Experiments tried with it showed that it reacted energetically with iodide 
of benzyl, giving tetrabenzyl-phosphonium iodide, also that it dissolved for the greater 
part in a hot solution of hydriodic acid of constant boiling point, and the solution after 
filtration through asbestos deposited a bulky crystalline precipitate. It was also found 
that ether dissolved the greater portion of the solidified oil, but left an insoluble 
crystalline solid. 

We therefore treated the whole of the solidified oil with ether, and filtered the solution 
from the solid crystalline mass remaining. The latter, when extracted with spirit and 
the solution crystallised, gave oxide of tribenzyl phosphine, which was identified by its 
melting point, 215°-215° # 5, and other properties. 

Now as all the operations had been conducted out of contact with air, this substance 
could not have been formed by the oxidation of tribenzyl phosphine. We shall consider 
its probable origin later on. 

We decided to precipitate the ethereal solution with gaseous hydriodic acid, and to 
examine the crystalline hydriodate which we anticipated would be that of the tertiary 
phosphine. In a preliminary experiment, we easily obtained a bulky crystalline precipi- 
tate. This, after washing with ether, was dissolved in hot solution of hydriodic acid, the 
solution filtered through asbestos, and allowed to cool, when a colourless bulky crystalline 
precipitate formed. This was thrown into a funnel plugged with glass wool, then drained 



562 PEOF. LETTS AND ME E. F. BLAKE ON 

on a plate of unglazed porcelain, and placed in a desiccator to dry, but it became 
so brown from separated iodine that it was not deemed advisable to analyse it. 
It was decomposed with a hot solution of caustic potash, when an oily liquid resulted, 
which was washed with water. Some of this, when exposed to the air and warmed, 
gradually became crystalline, and the crystals when recrystallised from alcohol had 
the appearance of oxide of tribenzyl phosphine. Another portion, when mixed with 
crystallised iodide of benzyl, evolved heat, and the product when recrystallised from 
alcohol separated in the characteristic form of tetrabenzyl phosphonium iodide, and had 
the melting point 189° C. After these preliminary experiments, it was decided to 
precipitate the remainder of the ethereal solution with hydriodic acid gas ; but as the 
ethereal solution contained a little water, fractional precipitation was resorted to, as we 
considered it probable that the first portions of the precipitated hydriodate would carry 
down the whole of the water. This surmise was fully borne out by experiment, for 
on passing hydriodic acid through the solution a quantity of a slimy white precipitate 
separated, and on filtering from this and passing hydriodic acid through the filtrate, a 
pure white flocculent precipitate was obtained, which was (after the solution had been 
saturated) thoroughly washed with dry ether, and then dried in vacuo. On analysis, the 
following numbers were obtained : — 



Analysis. 

(1) 1-0223 


required 26 c.c ^ 


AgN0 3 


= 03302 I = 32-29 per cent. 


(2) 0-5028 


12-3 c.c £ 


)> 


= 0156211 = 3106 




Found. 
(1) 31-06 




Calculated for 


Iodine, 


(C 7 H 7 ) 2 HP.HI. (C 7 H 7 ) 3 P.HI 
3713 29-39 


>> 


(2) 32-29 




... 



The above analyses indicated, in our opinion, that the compound was a mixture of 
the hydriodates of dibenzyl phosphine and tribenzyl phosphine, a view of its composi- 
tion which was rendered the more probable from the experiments described on p. 568 on 
the action of iodide of benzyl on the primary phosphine, which were made almost at the 
same time. We therefore decided to obtain a fresh and larger quantity of the 
hydriodates. 

Experiment 3. — The contents of twelve sealed tubes were treated exactly as in 
experiment 2, and an ethereal solution obtained of the " solidified oil." 

The bulk of this was treated with gaseous hydriodic acid, when a large quantity 
of a white precipitate was formed. An attempt was made to saturate the solution 
with hydriodic acid, but unfortunately the ether itself was rapidly attacked, yielding 
water and iodide of ethyl, with the result that the hydriodates of the phosphines 
were partly dissolved and a slimy mass obtained, which we could not dry nor do 
anything with. 

Some of the ethereal solution of the cake of crude phosphines was mixed with a 
filtered solution of sulphur in bisulphide of carbon, when much heat was evolved, and 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 563 

a dense white precipitate formed. This was thrown into a filter, and then washed about 

six times with ether. Then, to remove any sulphur, it was boiled with chloroform (in 

which very little dissolved), thrown into a filter, and washed with chloroform. It was 

then dried and analysed. 

Analysis. 

0-4248 eave i 11443 C0 * = °' 312082 C = 73 " 46 P er cent " 
4248 gave } . 2514 H2O==0 . 027904 H= 6 . 57 

Obtained. Calculated for (C 7 H 7 ) 3 PS 

Carbon, . . . 7346 7500 

Hydrogen, . . . 6"57 625 

The compound was evidently not pure, but as it was insoluble in alcohol, and only 

sparingly soluble in chloroform, no method of purification suggested itself immediately. 

Eventually it was found that it would be recrystallised from boiling glacial acetic acid, 

in which it is slightly soluble, crystallising as the solution cools in tufts of needles. 

Some of the product was thus recrystallised several times and analysed. 

Analysis. 

017Q2 aav J0-1089H 2 O-00121 H = 675 per cent. 
g \ 04883 CO, = 01331727 C - 7431 





Obtained. 


Calculated for (C 7 H 7 ) 3 PS 


Carbon, 


74-31 


7500 


Hydrogen, . 


6-75 


6-25 



Sulphide of tribenzyl phosphine fuses at 276° (corr.), and darkens on exposure to 
light. It is evidently difficult to purify. 

Experiment 4. — The contents of eleven sealed tubes were treated as before, and 
an ethereal solution obtained of the " solidified oil." This was filtered and distilled to 
dryness in a current of carbonic anhydride. The residue consisted of about 50 c.c. 
of a yellow fluorescent liquid which was very viscous. It was mixed with about 500 c.c. 
of pure dry benzol, in which it easily dissolved on agitation. On leaving this mixture 
undisturbed (the flask being filled with carbonic anhydride to prevent oxidation) a bulky 
crystalline precipitate formed after some hours. This we at first took to be oxide of 
tribenzyl phosphine, but subsequent investigations proved it to be a totally different 
substance. We shall refer to it later as the " Insoluble crystalline body." 

The benzol solution filtered from it was saturated with dry hydriodic acid, when an 
abundant white crystalline precipitate of the phosphine hydriodates was formed. This 
was collected on a cloth filter, and washed two or three times with dry ether. The mass 
drained as far as possible from adhering ether by pressure between filter paper weighed 
88 grms. 

It was next boiled in a stream of hydrogen with solution of caustic potash as long as 
either benzol or ether passed over ; the mixture allowed to cool, and the oily layer (after 
syphoning off the potash solution) mixed with ether. This latter did not, however, dissolve 
the whole product, but left a considerable quantity of a white slimy substance. The 



564 PROF. LETTS AND MR R. F. BLAKE ON 

ethereal solution was filtered off, and distilled to dryness in a current of hydrogen. 
The residue thus obtained weighed from 20-30 grms. 

It was submitted to distillation under diminished pressure, the distilling flask con- 
taining it being immersed in a bath of fusible metal. Average pressure during 
distillation 90 mm. ; temperature of fusible metal, 260°-300° C. ; temperature of liquid 
distilling over, 100°-180°. The distillation was stopped after about one-fifth of the 
total product had passed over, as signs of decomposition began to appear in the residue. 

The distillate was a colourless liquid. The residue when cold, a fluorescent, viscous 
(almost solid) mass. The distillate had the following properties : — 

(1) Exposed to the air it oxidised violently, and gave white fumes ; it grew so hot 
that a vessel containing it could not be touched. A syrup resulted. This had an acid 
reaction, and dissolved for the greater part in potash solution, leaving a slight residue, 
which was viscous. On adding hydrochloric acid to the potash solution, no precipitate of 
dibenzyl phosphinic acid was produced. On adding sulphate of copper to the neutralised 
potash solutions, no precipitate was produced in the cold, but on warming, the charac- 
teristic precipitate of monobenzyl phosphinite of copper was thrown down. 

(2) Treated with crystallised benzyl iodide, the mixture grew slightly warm, and from 
the product only a very small quantity of iodide of tetrabenzyl phosphonium could be 
obtained. 

These results show that monobenzyl phosphine was contained in the distillate, and 
probably only minute quantities of di- or tribenzyl phosphine. 

Our experiments so far, have undoubtedly shown that the semisolid mass which is left 
on digesting the sealed tube product with potash, &c, contains both the secondary and 
tertiary phosphines, but whence comes the primary phosphine ? Provisionally, we must 
assume that it is derived from the destructive distillation of the secondary phosphine. 

2(C 7 H 7 ) 2 P = (C 7 H 7 )PH 2 +(C 7 H 7 ) 3 P , 

and that the tertiary phosphine also is scarcely volatile without decomposition. 

It was at all events proved that we need not attempt to isolate the secondary and 
tertiary phosphine by distillation, and that some other method must be devised. 

Experiment 5. — The contents of twelve sealed tubes were treated as before with steam, 
and afterwards with potash, &c. The remaining semisolid cake of phosphines weighed 
about 97 grms. 

The cake was treated with a considerable quantity of ether, and the residue (A) 
reserved for examination. The filtered ethereal solution, together with the ethereal 
washings from the insoluble matter, were left undisturbed for about sixteen hours, when a 
good deal of crystalline matter (B) was precipitated. On filtering from this, and 
adding an additional quantity of ether to the solution, more insoluble crystalline matter 
was precipitated. The addition of ether was continued until no further precipitation 
occurred. 

The filtered ethereal solution was then boiled to dryness in a current of carbonic 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 565 

anhydride, when a yellowish liquid remained, which on cooling became semicrystalline. 
When perfectly dry it weighed 6 4 '5 grms. It was examined as follows : — 

(1) On adding ether, a good deal remained undissolved (probably "insoluble crystalline 
body"). 

(2) Six grms. of the crude product were mixed with 6 grms. of crystallised iodide of 
benzyl. On agitation the mixture grew excessively hot. When cold the mass was 
extracted with ether (to remove the excess of iodide of benzyl and any other impurities), 
then boiled with alcohol, and the solution allowed to crystallise. The resulting crystal- 
line mass of crude iodide of tetrabenzyl phosphonium weighed when dry 5 '5 grms., 
which with the quantity of substance operated upon correspond with 53 per cent, of 
tribenzyl phosphine. 

The crude iodide was purified by recrystallising it four times from alcohol. 

Analysis. 

0-3956 gave 01758 Agl = 0-095006 I = 24-01 per cent. 

Obtained. Calculated for (C 7 H 7 ) 4 PI 

Iodine, . . 24-01 2432 

To identify the iodide with absolute precision some of it was converted into acid 

sulphate (by boiling with diluted sulphuric acid), and an analysis made of the purified 

product. 

Analysis. 

05726 grms. gave 02679 BaS0 4 = 0*1103793 S0 4 = 19-27 per cent. 

Found. Calculated for (C 7 H 7 ) 4 P.HS0 4 

S0 4 . . . 19-27 1951 

(3) 5 '67 5 5 grms. were exposed to the air in a platinum dish, but no heat was de- 
veloped. Weak baryta solution was then added, and the mixture evaporated to dryness : 
redissolved several times with fresh quantities of water and evaporated to dryness, in 
order to ensure complete oxidation. The residue was then boiled with water, the solution 
filtered and mixed with hydrochloric acid, but no definite precipitate of dibenzyl phos- 
phinic acid could be obtained. 

However, on boiling the residue with caustic potash, then extracting with water and 
acidulating the solution with hydrochloric acid, a very bulky precipitate of dibenzyl 
phosphinic acid was obtained, which weighed when washed and dried 1"1300 grms.* The 
insoluble residue from the potash treatment was oily. It was extracted with ether several 
times, and the ethereal solution evaporated to dryness : the residue weighed 0*9401 grms. 
The residue from the ether treatment was boiled with alcohol, the solution filtered and eva- 
porated to dryness, when 2'4420 grms. of crude oxide of tribenzyl phosphine were obtained. 

There remained after the treatment with alcohol 1*1233 grms. of insoluble matter, 
which was proved to contain the substance we have called above " Insoluble crystalline 
body." 

* At first we thought it possible that the dibenzyl phosphinic acid might have owed its origin to the action of the 
potash solution on oxide of tribenzyl phosphine (see p. 558). But that this reaction does not occur when oxide of tri- 
benzyl phosphine is boiled with caustic potash solution, we proved by an actual experiment. 



11300 


grms. 


•9401 


» 


24420 


>> 


11223 


)> 



506 PROF. LETTS AND MR R. F. BLAKE ON 

Tabulating the above results, we obtained by oxidising 5 "6 75 5 grms. of the crude 
product, 

Crude dibenzyl phosphinic acid, ...... 

Substances soluble in ether, ...... 

Crude oxide of tribenzyl phosphine, ..... 

Substances insoluble in alcohol (containing " insoluble crystalline body"), 

56344 

The weight of dibenzyl phosphinic acid corresponds roughly with 20 per cent, of 
dibenzyl phosphine, and that of the oxide of tribenzyl phosphine with about 50 per 
cent, of tribenzyl phosphine. The experiment confirms former results, and shows that 
the product is a complex mixture, but is composed mainly of the two phosphines. 

Having thus definitely proved that the two phosphines are produced in Hofmann s 
reaction, we had next to devise a method for separating them from each other and for 
isolating them in the pure state. We had already tried fractional distillation, and had 
failed in accomplishing the desired object ; we had also tried, though without any large 
measure of success, to isolate the tertiary phosphine by crystallising it from alcohol. The 
experiment had indeed left us somewhat in doubt as to whether it is a solid or a viscous 
liquid. At the time, however, we had only a very crude and impure product at our dis- 
posal, containing a large quantity of oxide of tribenzyl phosphine ; whereas now we had a 
product, comparatively speaking, pure and free from that substance. As no other method 
could be devised for isolating the tertiary phosphine, we decided once more to see whether 
it was a solid substance capable of being crystallised from alcohol. Accordingly the 
remainder of the product, some 50 grms., was first treated with a somewhat large quantity 
of ether, when a considerable quantity of " insoluble crystalline body " was obtained. 
It was perfectly colourless, and weighed when dry 2 grms. The ethereal extract was 
filtered from this, and again diluted with ether, to see whether any more insoluble crystal- 
line matter would be precipitated, but the solution remained clear. The ethereal solution 
was then distilled to dryness in a current of carbonic anhydride, and alcohol added to the 
hot solution in such quantity that the mixture when boiled gave a clear solution, with only 
a few oily drops. This mixture (contained in a flask full of carbonic anhydride) deposited 
on cooling a good deal of a viscous liquid, and then after about twenty-four hours' rest a 
considerable quantity of minute colourless crystals separated, which when examined with 
a lens appeared to be thick quadratic plates. 

The following experiments were made with portions of the crystals : — 

(1) Warmed with crystallised iodide of benzyl, reaction occurred, and iodide of tetra- 
benzyl phosphonium was obtained. 

(2) The crystals seemed to be volatile without decomposition. 

(3) They were soluble in ether, but on exposure to air the solution grew turbid rapidly, 
and needle-shaped crystals separated out, which had the characteristic form of oxide of 
tribenzyl phosphine, and which gave its characteristic compound when treated with a 
solution of bromine in acetic acid. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 567 

(4) A quantity of the crystals were removed, quickly dried from adhering oily matter 
by pressure between filter-paper, then dissolved in bisulphide of carbon and mixed after 
filtration with a solution of sulphur in bisulphide of carbon. An immediate precipitate 
of fine needles occurred. These were filtered off, dried by pressure, then recrystallised 
from boiling glacial acetic acid, in which they are only sparingly soluble. On cooling 
nearly the whole of the compound separated in tufts of colourless fine needles. After a 
single recrystallisation from acetic acid, the corrected melting point of the substance was 
found to be 276° C. An organic analysis gave the following results : — 

Analysis. 

n \ n 9001 i °" 7935 CO. = 0-216409 C = 7408 per cent. 

(1) 2921 g av e{ . l704H2O==0 . 018933H= 6 . 4g n 

(2) 0-2654 ?ave J °' 7180 CO, = 0-195818 C = 73"78 per cent. 
(2) 2654 S ave ( . 1558 H 2 O = 00l7311 H= 6-52 „ 

Obtained. 

Calculated for (C 7 H 7 ) 3 PS 





(1) 


(2) 




Carbon, . 


7408 


73-78 


75-00 


Hydrogen, 


6-48 


6-52 


6-25 



We have already mentioned that sulphide of tribenzyl phosphine is very difficult to 
purify. Its melting point we had previously found to be 276. 

(5) Another quantity of the crystals, removed from the upper portions of the crystalline 

cake, and which looked pure and free from the oily matter, was very carefully removed 

and rapidly thrown into a filter placed in a funnel in a deep beaker full of carbonic 

anhydride. They were then several times washed with alcohol and the filter containing 

them rapidly transferred to a wide glass tube full of carbonic anhydride, which was 

then plugged with a perforated cork, and a stream of dry carbonic anhydride passed 

through it, while the tube was gently warmed. By these means the crystals were washed 

and dried with a minimum chance of oxidation. The dry crystals were then rapidly 

transferred to a small tube full of carbonic anhydride, and a combined organic analysis 

and phosphorus determination made by the method already alluded to. 

Analysis. 

C 1-0823 CO =0-295172 C = 81-22 per cent. 

0-3634 gave -< 02455 H 2 = 0027277H = 75 
I 01352 Mg 2 P 2 7 = 0037758 P = 1039 

Obtained. Calculated for (C 7 H 7 ) 3 P 
Carbon, . . . 8122 82'89 

Hydrogen, . . . 7 -50* 690 

Phosphorus, . . . 1039 10-19 

The numbers obtained in the above analysis are sufficiently close to those calculated 
for tribenzyl phosphine, to prove that we had at last succeeded in isolating that sub- 
stance, if not in an absolutely pure condition. 

In another experiment we still further identified the tertiary phosphine by obtaining 

* The bydrogen is no doubt too high, as the combustion tube was not warmed, and cold oxide of copper was used — 
both no doubt containing hygroscopic moisture. 

VOL. XXXV. PART II. (NO. 15). 4 Z 



568 PROF. LETTS AND MR R. F. BLAKE ON 

its selenium compound. To produce this substance some of the ethereal solution, obtained 

as we have described above, was mixed with bisulphide of carbon and freshly precipitated 

selenium. The selenium rapidly disappeared, and a white precipitate was formed. This 

was boiled with alcohol, in which it was insoluble, then dissolved in a large quantity of 

boiling glacial acetic acid, the solution filtered from excess of selenium, and allowed to 

cool, when beautiful silky needles separated, which were colourless. They melted at 

2 5 6° '5, and on analysis gave the following numbers : — 

Analysis. 

0-2786 f 0-1487 H 2 = 0016522 H= 593 per cent, 

gave <j 0680Q C q 2=0 . 185454 C =6656 „ 

Obtained. Calculated for (C 7 H 7 ) 3 PSe 

Carbon, .... 6656 6596 

Hydrogen, . . . 5'93 549 

The selenium compound decomposes slowly on exposure to light. 

Action of the Haloid Compounds of Benzyl on the Primary Phosphine. 

Action of Benzyl Iodide on Monobenzyl Phosphine. — We investigated this action 
chiefly with the view of obtaining dibenzyl phosphine, but also with the object of 
obtaining from it tribenzyl phosphine, by the further action of benzyl iodide. 

We naturally anticipated that if the reaction between the two substances occurred at 
all it would be of the usual kind, and would at once give rise to a simple product. Experi- 
ment has, however, shown that such is not the case, as the following results show. 

Experiment 1. — 6 grms. of crystallised benzyl iodide* were placed in a tube, which 
was then filled with carbonic anhydride, and 5 grms.t of the primary phosphine were 
added. The tube was sealed and the mixture agitated, when the benzyl iodide liquefied, 
and a turbid fluid resulted (probably from a trace of water in the benzyl iodide). After 
about ten minutes the mixture solidified to a snow-white crystalline mass, but no sensible 
rise of temperature occurred. The tube was opened after two days, and the hard solid 
mass which it contained broken up, and treated with washed and distilled ether. The 
ether was, however, not sufficiently dry, and as a consequence the mass became pasty from 
absorption of water. It was then thrown into a filter, which was placed in a desiccator 
in vacuo. It soon became hard and brown at the edges. 

It was found that boiling chloroform dissolved the product, and that crystals — pre- 
sumably of hydriodate of dibenzyl phosphine — separated on cooling. Accordingly, the 
whole was dissolved in boiling chloroform, but on cooling only a very small quantity of 
crystalline matter was obtained. 

Moreover, on evaporating off the chloroform, no solid product remained, but only a 
viscous liquid, from which we did not succeed in obtaining any definite substances. 

Experiment 2. — 12 grms. of the primary phosphine of boiling point 170°-190° C. 

* Prepared by a very simple method discovered by us, namely, by saturating benzyl alcohol with dry hydriodic acid 
and washing the product with water when it solidifies. The yield appears to be quantitative. 

t A large excess of the primary phosphine was therefore employed, as the molecular weight of C-H-I is 218, while 
that of ( ' 7 H-PH 2 is 124 ; so that, roughly speaking, 2 parts of the former is required for 1 of the latter. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 569 

were mixed with 70 c.c. of absolutely dry ether.* 22 grms. of pure iodide of benzyl 
were dissolved in 150 c.c. of dry ether, and the two solutions mixed. A colourless 
liquid resulted, which, with the exception of growing slightly turbid, did not suffer any 
noticeable change after twenty-four hours. 

The ether was therefore distilled off. When most of it had passed over, the viscous 
residue began to crystallise rapidly, and a brisk action appeared to occur, a considerable 
quantity of hydriodic acid coming off. The flask was removed from the water-bath in 
which it had been heated, and when cold a quantity of dry ether was added to its contents, 
which consisted partly of crystalline matter, but chiefly of a brown viscous mass. The 
latter, in contact with the ether, gradually became white and crystalline, eventually 
completely so. This crystalline product was pounded up, and carefully washed with 
dry ether. 

Some of the product thus purified was boiled with water, and yielded an oily liquid, 
which on cooling partly solidified. This was treated with ether, when a white crystalline 
solid remained, and on crystallising it from alcohol it had the characteristic form of 
iodide of tetrabenzyl phosphonium. The whole of the product was treated with water and 
steam in a current of hydrogen to decompose the hydriodates and to liberate the phos- 
phines. No primary phosphine passed over. The water when cold was poured off, 
and the semisolid mass again steamed with fresh water, and this poured off when cold. 
Next the mass was shaken with ether and the solution filtered from undissolved crystal- 
line matter. The latter, on recrystallisation, had the characteristic form of oxide of 
tribenzyl phosphine. 

The ethereal solution on keeping deposited a good deal of crystalline matter. After 
a day or two it was decanted from this, and boiled to dryness in a stream of 
hydrogen. A colourless liquid remained, which gradually deposited crystals. The whole 
was distilled in vacuo, the distilling flask being immersed in a fusible metal bath. 
Pressure about 190 mm.; temperature of metal bath, 260°-270°. A fair quantity of liquid 
distilled, its vapour having a temperature during distillation of 120°-125°. The residue 
decomposed somewhat. (We discuss the properties of this distillate on p. 570.) 

Experiment 3. — 22 grms. of crystallised iodide of benzyl were placed in a flask full 

of carbonic anhydride, and 12 grms. of the primary phosphine (boiling point 170°-190°) 

added. The iodide of benzyl dissolved on agitation, then action gradually set in, and after 

some hours the mixture became solid. After about twenty-four hours, some of the solid 

product was removed, pounded up in a mortar, and washed six times by decantation 

with dry ether, until a drop of the washings left no appreciable residue on evaporation. 

It was then placed in vacuo, and formed when dry a colourless powder. 

Analysis. 

0-4428 gave 03032 Agl =0163857 1 = 3700 per cent. 

Obtained. Calculated for (C 7 H 7 ) 3 PH.HI 

Iodine, . . . 370 3713 

* Distilled from phosphoric anhydride. 



570 PROF. LETTS AND MR R. F. BLAKE ON 

The rest of the product was washed with dry ether, and then decomposed by potash 
in a flask through which a stream of hydrogen flowed. A current of steam was then passed 
to complete the decomposition, and to drive off any volatile matters. The aqueous 
solution was then decanted, and the oily liquid reserved for examination. 

(1) On mixing some of this liquid with iodide of benzyl in excess, much heat was 
disengaged, and on cooling a viscous mass resulted. This boiled with caustic potash 
gave a solid substance, and on crystallising the latter from alcohol two sets of crystals 
were obtained, one having the characteristic form of iodide of tetrabenzyl phosphonium, 
the other that of tribenzyl phosphine oxide. 

(2) The rest of the oily liquid was left for some time and became semisolid. On 
treating this with ether, a white crystalline solid remained undissolved, which was 
apparently oxide of tribenzyl phosphine. The ethereal solution filtered from this was 
distilled off in a stream of hydrogen, and left about 10 grms. of a colourless liquid, which 
was distilled in vacuo. 

It behaved in exactly the same manner as described with the corresponding liquid 
obtained in Experiment 2 ; that is to say a little liquid distilled over — say about one-third 
of the original quantity operated on — and the residue then showed unequivocable signs 
of decomposition with separation of red phosphorus. The distillate was mixed with the 
corresponding distillate described under Experiment 2, and the mixture submitted to an 
oxidation experiment conducted quantitatively. 

1*3995 grms. were weighed out in a sealed tube and transferred to a beaker. Dense 
white fumes were produced at first, and a syrupy liquid resulted. After twelve hours, 
baryta solution was added, and the mixture evaporated to dryness several times, with 
addition of water. The aqueous solution was then filtered and mixed with hydrochloric 
acid, when a crystalline precipitate of dibenzyl phosphinic acid was produced, which weighed 
when dry 0'0912 grms. The residue insoluble in water was treated with a little hydro- 
chloric acid to remove carbonate of barium, then washed and dried. Its weight amounted 
to 0*1098 grms. It consisted mainly at all events of oxide of tribenzyl phosphine.* 

The filtrate from the dibenzyl phosphinic acid and the hydrochloric acid washings 
from the oxide of tribenzyl phosphine were evaporated on a water-bath to drive off 
excess of hydrochloric acid, the residue neutralised with baryta, and again evaporated to 
dryness, and extracted with alcohol. The alcoholic extract when dry weighed 0*3786 
grms. This residue gave all the characteristic reactions of benzyl phosphinite of barium, 
and was considered to be that substance. The residue from the alcoholic extraction was 
dissolved in water, and the solution boiled, when the characteristic crystals of benzyl 
phosphinate of barium were precipitated. 

These results prove that the distilled liquid was chiefly monobenzyl phosphine 
mixed with smaller quantities of dibenzyl and tribenzyl phosphine. 

* The dibenzyl phosphinic acid was identified by its melting point (191 -5-192 ) after recrystallisation. The oxide 
of tribenzyl phosphine by its characteristic crystalline form and by the production of its crystalline compound with 
bromine in acetic acid. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 571 

The percentage weights of the oxidised products referred to the weights of the original 
liquid are as follows : — 
From — 

/P IT \PTT J ( C 7 H 7) PH 2°2 .... 18-8 

^ 7 n 7 ;r±i 2 .... { (C 7 H 7 )PH 2 3 .... ? 

(C 7 H 7 ) 2 PH .... (C 7 H 7 ) 2 PH0 2 653 

(C 7 H 7 ) 3 P .... (C 7 H 7 ) 3 PO 7-84 

Experiment 4. — 22 grms. of benzyl iodide were placed in a flask, and the latter filled 
with carbonic anhydride, and 12 grms. of the phosphine (several times rectified and 
boiling from 175°- 185° C.) were then added. The mixture was very slightly warmed, so 
as to melt the iodide of benzyl, and then agitated to ensure thorough mixing. 

Although the crystallised benzyl iodide had been dried as far as possible by pressure 
between filter-paper, it must have been wet, as small drops of water were distinctly 
visible in the mixture. 

The flask containing it was loosely corked, and set aside. In about ten minutes 
solidification commenced and proceeded gradually, but suddenly, after the lapse of about 
twenty minutes more, the partially solidified mass grew hot, and emitted a cloud of 
hydriodic acid. It was left at rest for some hours, and had by that time become brown 
at the edges. It was now removed from the flask, pounded up, and repeatedly washed 
with ether until nothing further was extracted. The product consisted of a snow-white 
powder. It was dried by pressure between filter-paper and finally in vacua. 

The following experiments were made with it : — 

(1) 4 '69 3 grms. were mixed with caustic potash solution, and the mixture evaporated 
to dryness on a water bath ; redissolved in water, again evaporated, and so on two or three 
times to ensure complete oxidation. The residue was then treated with water, and 
the solution filtered from the colourless insoluble matter which remained. 

The solution and washings were then precipitated with hydrochloric acid, when an 
abundant precipitate of dibenzyl phosphinic acid separated which, when washed and 
dried, weighed T304 grms. It was identified by its melting point, which after one 
crystallisation from alcohol, was found to be 193°. 

The insoluble matter was dried, and weighed 0*816 grms. After one crystallisation 
from alcohol, it had melting point of 211°, that of oxide of tribenzyl phosphine 
being 2 15° "5. 

Calculating from these numbers the quantities of dibenzyl phosphine hydriodate 
and tribenzyl phosphine hydriodate in the original product, the first would amount to 
about 38 per cent., the second to about 23 per cent. 

(2) We wished to ascertain with certainty whether the product contained, in addition 
to the secondary and tertiary compound, the quarternary iodide also. As the latter gives 
the oxide when boiled with potash solution, we decided to decompose the product with 
water. Accordingly, the remainder of it was placed in a small separating funnel, and 



572 PROF. LETTS AND MR R. F. BLAKE ON 

steamed for about an hour, while a current of carbonic anhydride passed at the same 
time to prevent oxidation. 

The oily liquid which resulted was washed two or three times with water, afterwards 
with a dilute and cold potash solution, and then shaken up with ether. A white powder 
remained undissolved, which was dissolved in a little hot alcohol. The solution deposited 
on cooling crystals having the characteristic form of tetrabenzyl phosphonium iodide. 
But we did not obtain sufficient for investigation. 

The results of the above experiments are curious and interesting. We think that only 
one conclusion can be drawn from them, namely, that the primary phosphine acts upon 
iodide of benzyl in the same manner that ammonia does upon an alkyl iodide, the three 
following reactions occurring : — ■ 

(1) (C 7 H 7 )PH 2 + C 7 H 7 I = (C 7 H 7 ) 2 HPHI 

(2) (C 7 H 7 )PH 2 + 2C 7 H 7 I = (C 7 H 7 ) 3 P.HI + HI 

(3) (C 7 H 7 )PH 2 + 3C 7 H 7 I - (C 7 H 7 ) 4 PI + 2HI . 

Thus confirming the results of our investigation of Hofmann's sealed tube reaction (see 
p. 566). 

The only other explanation, viz., that the primary phosphine used contained both the 
secondary and tertiary bases as impurities, is, we venture to think, untenable, for it had 
been repeatedly rectified, and the boiling points of di- and tri-benzyl phosphine, as we 
have shown, are at very high temperatures — so high indeed that they decompose almost 
completely when distilled. Moreover, in our experiments on the oxidation of the primary 
phosphine, no dibenzyl phosphinic acid nor tribenzyl phosphine oxide were obtained, 
and we should certainly have noticed them had they been produced. These views have 
been materially strengthened by our experiments on the action of chloride of benzyl on 
the primary phosphine. 

Action of Chloride of Benzyl on Monobenzyl Pliosphine — Experiment 1. — Five grms. 
of the primary phosphine (boiling point 175°-185°) were placed in a tube previously filled 
with carbonic anhydride, and 5 grms. of chloride of benzyl were added. The tube was 
then sealed and left for a night, but no signs of any reaction having occurred were 
apparent. The tube was then heated for about six hours at 130° C. Its contents then 
consisted of a somewhat viscous liquid, and on opening pressure was observed. 

A quantitative oxidation experiment was made with the product. 4 '3255 grms. 
were evaporated with water in a dish. It fumed, and the mass quickly grew crystal- 
line. The product was crystallised once from alcohol, and then had a melting point 
of 213°, and all the appearance of oxide of tribenzyl phosphine. It weighed 0*8373 grms. 
The alcoholic mother-liquors solidified when warmed with baryta solution, and the solid 
mass thus obtained appeared to consist of crude tribenzyl phosphine oxide also. It 
weighed when dry 0*8623 grms. The aqueous extracts and the baryta solution filtered 
from the crude oxide gave scarcely a trace of dibenzyl phosphinic acid, but they 
gave the reactions of benzyl phosphinous acid. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 573 

These results show that the product of the reaction yielded nearly 40 per cent, of 
its weight of tribenzyl phosphine oxide, and contained some benzyl phosphinous acid. # 

Experiment 2. — This experiment was made chiefly with the view to ascertaining 
whether the reaction gives rise to a phosphonium salt in addition to the tertiary phosphine. 
2 grms. of the phosphine and 2 grms. of chloride of benzyl were heated to 160° C. for five 
or six hours. The tube then contained a syrupy liquid and some amorphous phosphorus. 
On opening it there was very decided pressure from phosphuretted hydrogen (proved 
by its action on nitrate of silver). The viscous liquid fumed powerfully from escape of 
hydrochloric acid, and on adding water all the phenomena described in Experiment 1 
were reproduced. After thorough oxidation the solid mass was boiled with water, and 
the solution on cooling gave the characteristic long needles of chloride of tetrabenzyl 
phosphonium. These, after repeated recrystallisation from water, had a melting point 
of 228° "5 C. — exactly that of the pure chloride — and we estimated that over 1 grm. had 
been formed. 

The residue insoluble in water was extracted with potash, the solution filtered, and 
treated with hydrochloric acid, when no trace of dibenzyl phosphinic acid was precipitated. 
But on boiling the insoluble residue with alcohol plenty of tribenzyl phosphine oxide was 
obtained. 

The above experiments fully confirm the results of those which we made with benzyl 
iodide. They show that the action of the haloid derivatives of benzyl on the primary 
phosphine varies not only with the particular halogen present, but also with the conditions 
under which the action occurs. Thus, when benzyl iodide is employed at ordinary 
temperatures, the secondary phosphine is the chief product, while a considerable quantity 
of the tertiary phosphine is also formed, and only a small quantity of the phosphonium 
salt, but we suspect that at higher temperatures the proportions of these three products 
would not be the same. 

With benzyl chloride no action occurs at ordinary temperatures, but under the 
influence of heat a reaction takes place, the tertiary phosphine and the phosphonium 
chloride being formed, while if the temperature is high the phosphonium salt is the chief 
product, and no secondary phosphine is produced at all. 

Properties of Dibenzyl and Tribenzyl Phosphine. — Although we have not isolated 
the first of these in the pure state, and have only obtained the second in small quantities, 
we have so often examined a mixture of the two that we are enabled to state with 
probable correctness what their chief properties are. 

Dibenzyl Phosphine. — A liquid which cannot be distilled even in vacuo without 

considerable decomposition. This decomposition gives rise to the primary and tertiary 

bases, 

2(C 7 H 7 ) 2 HP^(C 7 H 7 )PH 2 + (C 7 H 7 ) 3 P . 

Dibenzyl phosphine combines with hydriodic acid, and probably with other hydracids, 

* The rest of the product was evaporated down with aqueous potash, and also yielded oxide of tribenzyl phosphine 
in abundance, some benzyl phosphinous acid, but scarcely a trace of dibenzyl phosphinic acid. 



574 PROF. LETTS AND MR R. F. BLAKE ON 

forming solid crystalline compounds which are decomposed by water. It forms an 
amorphous (?) orange-coloured chloroplatinate. 

It probably oxidises in the air, and there is some reason for believing that it then 
gives rise to an oxide C 14 H 15 PO (see p. 585). It certainly oxidises in contact with warm 
caustic potash solution and air, the product of its oxidation being then dibenzyl 
phosphinic acid, C 14 H 15 P0 2 . 

It is readily soluble in ether and benzol, but is sparingly soluble in alcohol. 

Tribenzyl Phosphine. — A solid crystalline substance of high boiling point. Possibly 
volatile in vacuo without decomposition or with only partial decomposition. It combines 
with hydracids, forming solid compounds which are decomposed by water. Its chloro- 
platinate is amorphous and buff- coloured. 

It oxidises in the air and fumes when warmed, spontaneous oxidation occurring very 
readily, the product being the oxide (C 7 H 7 ) 3 PO. It also combines energetically with 
sulphur and selenium to give colourless crystalline compounds, which are insoluble or 
very sparingly soluble in water, alcohol, ether, bisulphide of carbon, and chloroform, but 
are soluble in boiling glacial acetic acid, from which they may be crystallised. Both 
compounds decompose on exposure to light, and have the formulae (C 7 H 7 ) 3 S and (C 7 H 7 ) 3 Se, 
respectively. Tribenzyl phosphine combines energetically with benzyl iodide to give 
tetrabenzyl phosphoDium iodide. 

We may mention that attempts have been made to obtain tribenzyl phosphine by 
several other methods, but without success. Among them were — 

(1) The action of chloride of benzyl on phosphide of sodium.* 

(2) The action of benzyl alcohol on phosphonium iodide in a sealed tube. This 
action has also been studied by LEDERMANN.t Apparently only the phosphonium iodide 
is produced. 

(3) The action of sodium on a mixture of chloride of phosphorus and benzyl bromide. 
It was by a corresponding method that Michaelis| obtained triphenyl phosphine. We 
found that no action occurred. 

(4) The action of sodium on tetrabenzyl phosphonium chloride. § 

(5) The action of heat on chloride of tetrabenzyl phosphonium. || 

Compounds of Tetrabenzyl Phosphonium. 

Salts of tetrabenzyl phosphonium have been obtained by the following methods : — 
The chloride, (1) by the action of chloride of benzyl on phosphide of sodium (Letts 

and Collie).! The method is an excellent one, being easily carried out, and giving a 

good yield of the compound. 

(2) It is also formed by the action of chloride of benzyl on the primary phosphine 

(p. 573). 

* Letts and Collie, these Transactions, xxx. part 1. t Ledermann, Berichte, xxi. (1888) 405a. 

I Michaelis, Berichte, xv. (1882) 801a. § Letts and Collie, loc. cit. || Letts and Collie, loc. cit. 

1 Letts and Collie, loc. cit., p. 181. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 575 

The iodide (1), by the action of phosphonium iodide on benzyl alcohol (Ledermann),* 

(2) By the union of tribenzyl phospbine with iodide of benzyl (p. 566). 

(3) In small quantities by the action of iodide of benzyl on the primary phosphine. 
The salts of the phosphonium are beautifully crystalline, and are readily obtained in a 

state of purity. All are soluble in alcohol and many in water ; they are insoluble in 
ether. When heated they decompose as a rule, and in many cases give rise to perfectly 
definite products. 

Chloride (C 7 H 7 ) 4 PC1,2H 2 0. — Colourless salt crystallising from boiling water in long 
needles. It is sparingly soluble in cold water (100 c.c. of cold water dissolve about 
0"35 grms.) and almost insoluble in dilute hydrochloric acid or common salt solution 
(traces may be detected in an aqueous solution by adding these reagents). The dried 
salt dissolves readily in alcohol, and the solution on slow evaporation yields beautiful 
colourless rhombic crystals of considerable size, which are anhydrous. The dried salt also 
dissolves readily in chloroform, and on spontaneous evaporation large crystals are also 
obtained, which grow opaque in the air from loss of chloroform. The crystals thus 
obtained contain one molecule of " chloroform of crystallisation." The anhydrous salt fuse3 
at 228°'5 (corr.). 

The chloride combines with chloride of platinum and some other metallic chlorides, 
giving insoluble or separately soluble compounds. 

Chloroplatinate, 2(C 7 H 7 ) 4 PCl,PtCl 4 (Letts and Collte). — 'Obtained in minute orange- 
coloured crystals by mixing alcoholic solutions of the two chlorides. Almost insoluble. 

Double Salt with Mercuric Chloride (Ledermann) — (C 7 H 7 ) 4 PCl,HgCl 2 ,H 2 0. — In- 
soluble precipitate obtained like the chloroplatinate. 

Double Salt with Stannic Chloride, 2(C 7 H 7 ) 4 PCl,SnCl 4 (Ledermann). — Sparingly 
soluble crystalline compound obtained like the preceding salts. 

Bromide, (C 7 H 7 ) 4 PBr, resembles the chloride, but is less soluble in water. It dissolves 
readily in alcohol, and crystallises from a mixture of alcohol and water in long silky 
needles. Melting point, 2 16°-217° C. (uncorrected). 

Iodide, (C 7 H 7 ) 4 PI, is almost insoluble in water, and only sparingly soluble in alcohol. 
From a hot alcoholic solution it crystallises in small rhombic crystals.t 

All the haloid salts appear to be decomposed completely when boiled with alkalies 
into the tertiary phosphine oxide and toluol, 

(C 7 H 7 ) 4 PX + MOH = MX + (C, H 7 ) 3 PO + C 7 H 8 . 

Acid Sulphate, (C 7 H 7 ) 4 PHS0 4 , is most readily obtained by warming the chloride with 
oil of vitriol on a water-bath until hydrochloric acid ceases to come oif, and then 
crystallising the product once or twice from boiling water. It is rather more soluble 
than the chloride, and crystallises in small rhombic or triclinic crystals. When boiled 

* Ledermann, Berichte, xxi. (1888) 405a. 

t We are of opinion that when this compound is crystallised from alcohol, the crystals have the formula 
2(C 7 H 7 ) 4 PI,C 2 H 6 0. 

VOL. XXXV. PART II. (NO. 15). 5 A 



,")76 PROF. LETTS AND MR R. F. BLAKE ON 

with caustic baryta it is partly converted into the hydrate and partly into the tertiary 
phosphine oxide, and toluol, 

(1) (C 7 H 7 ) 4 PHS0 4 + Ba(OH) 2 = (C 7 H 7 ) 4 POH + BaS0 4 + H 2 

(2) (C 7 H 7 ) 4 PHS0 4 +Ba(OH) 2 =(C 7 H 7 ) 3 PO + BaS0 4 +2H 2 + C 7 H 8 . 

Treated with carbonate of barium, the hydrate only is obtained — 
(C 7 H 7 ) 4 PHS0 4 + BaC0 3 = (C 7 H 7 ) 4 POH + BaS0 4 + C0 2 . 

Normal Sulphate, {(C 7 H 7 ) 4 P} 2 S0 4 , 6H 2 0. — Obtained by decomposing a solution of 
the chloride with sulphate of silver. It crystallises from a hot concentrated solution in 
large rhombic plates. 

Nitrate, (C 7 H 7 ) 4 PN0 3 . — Long silky needles, soluble in water. 

Chromate, {(C 7 H 7 ) 4 P} 2 Cr0 4 ? — Obtained by double decomposition from the chloride 
and chromate of silver. Small lemon yellow plates. 

Acetate, (C 7 H 7 ) 4 P(C 2 H 3 2 ). — Very soluble in water. 

Chlorate, (C 7 H 7 ) 4 P(C10 3 ). — Crystallises in long needles from a moderately con- 
centrated solution. When heated it puffs. 

Picrate, (C 7 H 7 ) 4 P{C 6 H 2 (N0 2 ) 3 0} (Ledermann). — Obtained by mixing alcoholic solu- 
tions of the iodide and picric acid ; the compound separating as the solution cooled in 
beautiful yellow crystals. 

Hydrate, (C 7 H 7 ) 4 P(OH). — Obtained best by the action of carbonate of barium on a 
solution of the acid sulphate. Very soluble even in cold water, and crystallises by slow 
evaporation in beautiful rhombohedral plates sometimes half an inch long. The crystals 
are transparent and highly refractive. The hydrate dissolves readily in alcohol, and the 
crystals obtained from the solution contain alcohol of crystallisation, their formula being 
(C 7 H 7 ) 4 P(OH),C 2 H 6 0. Solutions of the base have an alkaline reaction. They neutralise 
acids with formation of salts of the phosphomum. When heated the hydrate decomposes 
into the tertiary oxide and toluol, 

(C 7 H 7 ) 4 POH = (C 7 H 7 ) 3 PO + C 7 H 8 . 

The Bye Products formed in Hofmann's Sealed Tube Eeaction. 

In separating the benzyl phosphines from the products of Hofmann's sealed tube 
reaction we obtained a considerable number of bye products, among which were the 
following : — 

(A). A crystalline substance precipitated on addition of hydrochloric acid to the potash 
solution which had been employed to extract the viscous mass containing tribenzyl 
phosphine, &c. 

(B). A crystalline substance which separated spontaneously from the aqueous solution 
obtained by treating the contents of the sealed tubes with water. 

((7). A crystalline zinc salt, also contained in the same aqueous solution. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 577 

(.D). A crystalline solid of low melting point remaining in the alcoholic mother-liquors 
from which oxide of tribenzyl phosphine had crystallised out. (This substance we call 
"Crystalline Oil".) 

We obtained all of these bye products in our earlier experiments, in which after 
steaming the contents of the sealed tubes with water to liberate the primary phosphine, 
the residual viscous insoluble mass was boiled with aqueous potash. 

In our later investigations other bye products were obtained, but we shall first discuss 
the composition of the first three we have just enumerated. 

A. Crystalline Substance precipitated by Hydrochloric Acid from the Potash Solution 
used for extracting the viscous mass containing Tribenzyl Phosphine, &c. — This body 
was precipitated in crystalline flocks. It was very sparingly soluble in cold water, 
rather more so in hot water, and crystallised from a boiling aqueous solution in indistinct 
leaflets. That the substance had acid properties was proved by the readiness with which 
it dissolved in potash and baryta solutions. A slight residue was, however, left in 
both cases, and indeed an impurity appeared to be present which was extremely difficult 
to eliminate, as the following analyses show. 

Analysis. 

Obtained. 



I. 
i, . . 59-69 
jen, . . 5-80 


A. 
II. 

66-30 
630 


III. 

6706 

6-50 


Calculated for 
(C 7 H 7 ) 2 P0 2 H 
68-29 
609 



(1) Crude product washed with water. 

(2) Precipitated from a solution of the crude product in baryta, and carefully washed. 

(3) Several times dissolved in baryta and reprecipitated by hydrochloric acid, then recrystallised from alcohol and water. 

The identity of the substance was, however, fully established by analyses of some of 
its salts. 

Barium Salt. — Obtained by dissolving the crude product in baryta water, and 
subsequent precipitation of the excess of baryta by a stream of carbonic anhydride. 
The salt crystallised from the highly concentrated solution in tufts of thin plates. 

Analysis. 

Obtained. ,-, , ■, . , c 

Calculated tor 



L ~~ iT {(C 7 H 7 ) 2 P0 2 } 2 Ba 

Barium (in salt dried at 110), 22-2 21'8 21-8 

Zinc Salt. — Obtained as a white amorphous precipitate on adding acetate of zinc to a 
solution of the barium salt. 
Analysis. 

,-.,, ■ i Calculated for 

Obtained. { (C 7 H 7 ) 2 P0 2 } 2 Zn 

Zinc, . . . . 121 11-7 

Silver Salt. — Obtained by adding a strong solution of nitrate of silver to a solution 
of the acid in alcohol, when the salt separated in thin colourless needles. 



57$ PROF. LETTS AND MR R. F. BLAKE ON 

Analysis. 

,-.,,- , Calculated for 

Obtained. (C 7 H 7 ) 2 P0 2 Ag 

Silver, .... 30-1 306 

Subsequently we obtained dibenzyl phosphinic acid by other methods, and satisfied 
ourselves that it was identical with the substance in question. 

B. Solid Substance which separated spontaneously from the Aqueous Solution obtained 
by treating the contents of the Sealed Tubes with Water. — This body separated from the 
solution as a white powder stained brown by free iodine. On boiling with dilute spirit 
it dissolved, and the solution on cooling deposited thin colourless plates, which when 
dried had the lustre of mother-of-pearl. An analysis was not made, as the melting point 
(192° corr.) and other properties proved sufficiently that the substance was dibenzyl 
phosphinic acid. It may be mentioned that this body is also produced when chloride of 
benzyl and phosphonium iodide are heated alone in a sealed tube, the product of action 
being; afterwards treated with water. 

C. Crystalline Zinc Salt contained in the Aqueous Solution obtained by treating 
the contents of the Sealed Tubes with Water. — This salt was obtained as follows : — The 
aqueous solution filtered from B (dibenzyl phosphinic acid) was mixed with excess 
of acetate of lead (in order to precipitate hydriodic acid), then after filtering from 
iodide of lead saturated with sulphuretted hydrogen, again filtered and evaporated to 
small volume. "When most of the acetic acid had volatilised a zinc salt crystallised out in 
small nodules. Determinations of zinc and water indicated that this salt was benzyl 
phosphinate of zinc, but it was not obtained in sufficient quantity for further examina- 
tion. 

Analysis. 

Obtained. Calculated for (C 7 H 7 )P0 3 Zn,H 2 

Zinc, .... 257 257 

Water, .... 6*9 71 

In a later series of experiments a different method of treatment was employed, with 
the object of obtaining all the bye products of the reaction. 

The contents of the sealed tubes were steamed with water as already described, and 
the aqueous solution decanted. The brown viscous mass remaining was then steamed a 
second time with water, and the solution decanted and mixed with the first. The 
viscous mass, now fairly free from soluble zinc salts and hydracids, was repeatedly 
extracted with alcohol, when a considerable quantity was dissolved, leaving, however, 
a resinous mass coloured with red phosphorus. This latter was repeatedly boiled with 
baryta, and then extracted with chloroform, which dissolved a resin having a strong green 
fluorescence, and containing phosphorus. We have not succeeded in ascertaining its 
nature. The alcoholic extracts were thrown into a large volume of caustic baryta solution 
and boiled for a considerable time, and this treatment repeated with what remained 
undissolved. The resinous matter at first precipitated became crystalline during the 
ebullition, and eventually the crystals were found to consist of oxide of tribenzyl phos- 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 579 

phine. The baryta extracts filtered off were mixed with those which had been employed in 
extracting the resin which was insoluble in alcohol, and the mixture deposited on cooling 
a small quantity of oxide of tribenzyl phosphine. This was filtered off, and excess of 
baryta removed from the solution by a current of carbonic anhydride. The filtered 
solution gave on concentration a considerable quantity of '''crystalline oil," which was 
filtered off. To the filtrate from this, hydrochloric acid was added, when an oily liquid 
was precipitated which rapidly solidified to a crystalline mass. The latter was recrys- 
tallised from alcohol, and was identified by its melting point (190° corr.), appearance, and 
by an analysis of its barium salt as dibenzyl phosphinic acid. 

Analysis of Barium Salt. 

Obtained. Calculated for {(C 7 H 7 ) 2 P0 2 } 2 Ba,8H 2 

Barium, . . . 16-99 1777 

Water, . . . 1858 1867 

The mother-liquors from which the acid had been separated by hydrochloric acid were 
mixed with sulphuric acid, filtered from the precipitated sulphate of barium, and evaporated 
on a water-bath until all hydrochloric and hydriodic acid had volatilised. The residue was 
diluted, neutralised with chalk, filtered and evaporated to small bulk, when a sparingly 
soluble calcium salt separated in crystalline crusts. These, after being washed first with 
water and then with spirit, were dried at 110° and a determination of calcium made. 

Analysis. 

Obtained. Calculated for {(C 7 H r )HP0 2 } 2 Ca 

Calcium, .... 1108 11-42 

The properties of the acid from which the calcium salt was prepared, such as the 
production of the primary phosphine on heating, and the characteristic reaction with a 
copper salt, further identified it as benzyl phosphinous acid. 

D. A Crystalline Substance of low Melting Point (" Crystalline Oil "). — This 
substance was contained in the viscous mass which remained after treating the product 
of the sealed tube reaction with aqueous caustic potash, or, if alcoholic potash is employed, 
it passes into solution. 

It dissolved sparingly in water, and separated out from a hot solution in oily drops, 
which eventually solidified. 

It was produced in only small quantity — the largest amount which we have ever 
obtained being about 9 grms. from the contents of twenty-four tubes, representing 
288 grms. of chloride of benzyl. 

The investigation of this substance has given us a great deal of trouble, and we are 
still uncertain as to its composition. The following analyses were made by one of us and 
W. Wheeler : # — 

(I.) Crude substance washed with ether. 



0-2741 i °' 6922 C0 2 = 018879 C =68-88 per cent. 

gaVG (01650 H 2 = 001833 H= 6-68 

* Proceedings, 1887, p. 82. 



580 PROF. LETTS AND MR R. F. BLAKE ON 

(II.) The crude substance was dissolved in boiling dilute baryta solution, the baryta 
was then precipitated by carbonic anhydride, and the substance obtained from the 
concentrated solution. 

0-3355 ^ve J 0-8790 C0 2 = 0'2397 C =71-45 per cent. 
*• I 02072 H 2 = 002302 H= 686 

(III.) The crude substance was recrystallised twice from water. 

02474 e-ave i °' 647 C0 2 =01764 C =7130 per cent. 
2474 gave { . 1553 H 2 o = . 01725 H== 6 - 97 

(IV.) Another specimen also recrystallised twice from water. 

04412 o-ave J H606 CO, =0-3165 C =71-73 per cent. 
g (02677 H 2 = 002974 H= 6-74 

The next set of analyses were made by us after the investigation had been re-opened, 
and we were in possession of a fresh quantity of the substance. 
Its corrected melting point we now found to be lll o, 0. 
(V.) The specimen was twice recrystallised, and melted at lll°-0-lll°'l. 

„ QUfi / 08360 C0 2 = 0-2280 C = 7247 per cent. 

0314b gave | . 1923 H 2 = 0021366 H = 679 „ 

(VI.) The specimen was obtained from the mercuric chloride compound (vide infra), 
by decomposing it with sulphuretted hydrogen. In the combustion, oxide of copper 
unmixed with chromate of lead was used. 

ftt)/fQ „ f 0-6727 C0 2 = 0-183463 C = 73"67 per cent. 

2490 gave | . 1607 HaO = 0<)17855 H= 7 . 17 ; 

(VII.) The same specimen as was employed for No. V. analysis. The combustion 

was made with pure oxide of copper, and a phosphorus determination subsequently made 

(see p. 556). 

( 07826 C0 2 = 0-213436 C = 7316 per cent. 
02917 gave J 01912 H 2 =0021244 H= 7-28 
1 0-1481 Mg 2 P 2 7 = 0041361 P = 1418 

(VIII.) A different specimen recrystallised four times, and of constant melting point 
1 11°"0. The analysis was performed in the same way as VII. 







(086 
^0-21 


62C0 2 


= 0236236 C = 


= 72-75 


per cent. 






0-3247 


gave 


43 H 2 


= 0023811 H = 


= 7-33 


>) 










101623 Mg 2 P 2 7 = 0045327 P = 


= 13-95 


)> 










I. 


II. 


III. 


IV. 


V. 


VI. 


VII. 


VIII. 


Carbon, 


. 


68-88 


71-45 


71-30 


71-73 


72-47 


73-67 


7316 


72-75 


Hydrogen, 




668 


686 


697 


6-74 


6-79 


7-17 


7-28 


733 


Phosphorus, 


. 














1418 


1395 


Oxygen (by differe 


ace), 














5 38 


5-97 




10000 


10000 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 581 

The substance had the following properties : — 

(1) When heated most of it appeared to volatilise unchanged, but at the same time a 
slight odour of the primary phosphine was noticed. 

(2) Chloride of platinum gave no sparingly soluble nor crystalline compound. 

(3) Bromine vapour liquefied it, while the addition of liquid bromine caused the 
disengagement of heat and escape of hydrobromic acid. 

(4) Its aqueous solution when mixed with corrosive sublimate remained clear, but on 
warming the mixture a colourless amorphous precipitate was suddenly produced. 

(5) When iodide of cadmium was added to its hot aqueous solution oily drops separated 
out, and these gradually solidified, the resulting crystals being minute cubes or quadratic 
forms, very refractive to light. This compound was very sparingly soluble. 

(6) Iodide of zinc gave an oily precipitate, which refused to crystallise. 

(7) Stannic chloride and stannous chloride both gave white precipitates — probably con- 
sisting of minute oily drops. 

The iodide of cadmium compound was prepared several times, and analysed with the 
following results : — 

Analysis of Cadmium Iodide Compound. 

(I.) 

0-6448 aavei °" 3764 A S T =0 ' 2034 * = 31 ' 5 P er cent 
Ub448 gave | . 0796 CdO = . 0796 Cd = 12 . 3 m 

(II.) A different specimen. 

( 0-9687 Agl =05235 I =2865 per cent. 
1-8267 gave | . 2676 CdO = . 2341 Cd =12 - 8 1 „ 

n ,„ Q f 08559 CO 2 = 0-2334CO 2 =4074 „ 

oa/zy » | 01979 H 2 = 0021989 H = 3-84 „ 

(III.) A different specimen recrystallised from water. 

1-2994 { re( l uired 29 ' 8 c-c^AgNOj = 0-37846 I = 29-12 per cent, 
(gave 01783 CdO =0156012 Cd = 1200 

(IV.) A different specimen. The substance was burnt with pure oxide of copper, and 
the phosphorus determined by molybdate, &c. 

0-497 gave 0-1471 Mg 2 P 2 7 = 0041081 P = 826 per cent. 

(V.) The same specimen as IV., the analysis being performed in the same way. 

03109 gave 0-0919 Mg 2 P 2 7 = 0025665 P = 8"25 per cent. 

(VI.) The specimen was prepared from pure crystalline oil. 

n *m a f 08980 C0 9 = 0244909 C = 40-75 per cent. 

06010 gave J Q ^ Q H ^ = ( , 230 H = 382 „ 

11950 eave { 0-6845 Agl =0369921 I =30-95 „ 
g t 01550 CdO = 0135625 Cd = 11-34 



582 



PROF. LETTS AND MR R. F. BLAKE ON 



07947 


gave 


01209 CdO = 


= 01057875 Cd = 


= 13 31 


per cent 




•7340 


„ 


01020 CdO = 


= 008925 Cd = 


= 12-15 


>> 




10631 


f> 


06234 Agl = 


= 0336901 I = 


= 31-69 


>> 








I. 
12-3 


II. 
12-81 


III. IV. 
1200 


V. 




VI. 

A 


Cadmium, . 


11-34 


1331 12-M 


Iodine, 




315 


28-65 


2912 




3095 


31-69 ... 


Carbon, 






4074 






4075 




Hydrogen, . 




... 


384 




... 


3-82 


... 


Phosphorus, 






. . . . 


8-26 


8-25 






Oxygen (by difference), 


... 






... 


... 


... 



We had hoped that the analysis of the cadmium iodide compound would definitely 
settle the question of the composition of the crystalline oil, but such was unfortunately not 
the case, as we shall show presently in discussing the analytical results. As the quantity 
of crystalline oil at our disposal was very small, we were unable to study its properties 
exhaustively, and the difficulty of the investigation was thus increased. 

Eventually we decided to turn our attention to the corrosive sublimate compound 
which is very characteristic. It was easily obtained by mixing dilute aqueous solutions 
of the crystalline oil and corrosive sublimate and warming the mixture, when a bulky 
amorphous precipitate was produced, which had a faint blue tinge. This, when washed 
and dried, formed a light powder, which was extraordinarily electric. The compound was 
twice prepared and analysed, with the following results : — 

Analysis. 

0-8605 "">™ / 04397 Hg =51 -09 per cent. 



(1) 



gave 



(2) 05359 gave 



i 04397 Hg 
\ 0-3215 AgCl 

f 0-2387 Hg 

I 0-03763 CI (volumetrically) = 702 



0-07953 Cl= 924 

= 44 - 54 per cent. 



The ratio of mercury to chlorine is in both cases Hg : CI, for — 

(1) 200 : 35-5 :: 5109 : 907 

(2) 200 : 355 :: 44"54 : 790 

The mercury and chlorine were thus shown to be present as calomel, proving that 
part at least of the compound had been decomposed and probably oxidised. 

As the mercury compound employed in the two analyses were different preparations, 
it is evident that its composition is not constant, and in all probability it consists of 
calomel containing varying quantities of the original substance ; for, on decomposing some 
of it with sulphuretted hydrogen and concentrating the filtered solution, oily drops 
separated like the original body, and these gradually crystallised in its characteristic 
manner, and on analysis were found to contain the same proportions of carbon and 
hydrogen as the crystalline oil (No. VI. determination). The product of oxidation was 
easily found and identified in the mother-liquors of the calomel compound ; for, on con- 
centration and cooling, they deposited crystalline scales having the appearance of dibenzyl 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 583 

phosphinic acid. To obtain them as pure as possible, they were dissolved up by warming, 
and a current of sulphuretted hydrogen passed through the solution until the whole of 
the mercury was precipitated. On filtering the hot solution and allowing it to cool the 
scaly crystalline compound again came down, and a little more of it was obtained on 
evaporating its mother-liquors. The whole was then recrystallised from water, in which 
it was very sparingly soluble. 

It was identified as dibenzyl phosphinic acid by its melting point 193° (that of 
the pure acid being 192° C), by its solubility relations, by its crystalline form, and 
also, though not very satisfactorily, owing to the small quantity at our disposal, by a 
combustion. 



Analysis. 



( 01477 H 2 = 0016411 H= 619 per cent. 
0-647 gave -^ . 6507 C0 „ = 0-177463 C = 6704 „ 





Obtained. 


Calculated for (C 7 H 7 ) 2 HP0 2 


Carbon, 


67-04 


68-29 


Hydrogen, . 


619 


6-09 



We have been unable to go any further with our investigations, as the whole stock of 
crystalline oil at our disposal became exhausted, and, as we have before stated, we are 
still in doubt as to its composition. 

That it is an oxidised derivative is shown by the deficiency which remains when the 
percentages of the constituents both of it and of its cadmium iodide compound (which were 
actually determined) are added together. 

There can also be no doubt that it gives rise to dibenzyl phosphinic acid when 
oxidised, and therefore in all probability it contains the element of dibenzyl phos- 
phine. 

It will be noticed that of the eight analyses made of the crystalline oil itself, Nos. II., 
III., and IV. agree fairly well with each other, but differ from Nos. V., VI., VII., and 
VIII. , which also agree (though not so closely) with each other. Thus — 

Mean of II., III., and IV. Mean of V, VI., VII., and VIII. 

Carbon, .... 71*50 7301 

Hydrogen, . . . 6"87 7-14 

The first set of analyses agree with the formula C 13 H 15 PO, so far as the percentage amounts 
of carbon and hydrogen are concerned, while the second set agree in a similar manner with 
C 15 H 17 PO. The percentage amount of phosphorus, however, from Nos. VII. and VIII. 
agrees with the first formula. Thus — 

o 







Obtained. 


Calculated for 


Carbon, 
Hydrogen, . 
Phosphorus, 

VOL. XXXV. PART II. (NO. 


I. 
71-50 

6-87 

15). 


II. 

73-42 

7 23 

1406 


C 13 H 15 PO 
71-56 
6-88 
1442 


C 16 H 17 PO 

73-30 

697 

12-70 

5 B 



584 



PROF. LETTS AND MR R. F. BLAKE ON 



The mean results of the whole of the analyses are as follows :- 

Carbon, ........ 

Hydrogen, . . . . . . ' . 

Phosphorus, ....... 

Oxygen (by difference), . 



7336 
7-02 

1406 
656 



And the atomic ratios, taking the percentage of phosphorus as representing one atom 
of that elemeut — 

Next as regards the analysis of the cadmium iodide compound. It will be seen that 
the percentages of cadmium and iodine vary to a considerable extent in the different 
determinations. This must be attributed to the analytical difficulties we experienced in 
separating the two substances from such a complex body. On the other hand, the two 
phosphorus determinations and the two combustions agree with each other exceedingly 
well. The mean results are as follows : — 



Cadmium, 

Iodine, 

Carbon, 

Hydrogen, 

Phosphorus, 

Oxygen (by difference), 



1232 

3038 

40-74 

383 

8-25 

4-48 

10000 



and the atomic ratios, taking the percentage of phosphorus as representing 1 atom of 
that element, 

Thus the calculated atomic ratios of the crystalline oil and of the cadmium iodide 
compound are not very different, and so far the comparison is satisfactory. But, on the 
other hand, if we compare the percentage numbers obtained with those required for a 
cadmium iodide compound of either C 13 H 15 PO or C 15 H 17 PO, a total discordance is seen. 
Thus— 

Calculated for 



Cadmium, 


(C 13 H 16 PO) a CdI 2 
13-96 


(C 16 H I7 PO) 2 CdI 2 
131 


Iodine, 


. 


3167 




29-7 


Carbon, . 




3890 




421 


Hydrogen, 
Phosphorus, 




3-70 
7-73 




39 

7-26 



In fact, no formula can be devised which will agree exactly with the analytical results 
obtained with both the crystalline oil and its cadmium iodide compound. The numbers, 
however, calculated for the formula Ci 4 H ]5 PO approximate pretty closely to those 
obtained. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 585 



Thus- 





Calculated for 




C 14 H 15 PO 




(Cu 


H l5 P0) 2 CdI 2 


... 






13-56 
30-85 


7304 






4068 


6-52 






363 


13-48 






750 



Cadmium, 
Iodine, 
Carbon, . 
Hydrogen, 
Phosphorus, 

This would be the formula of an oxide of dibenzyl phosphine, 

C 7 H 7 . 

C 7 H 7 ->P = 0, 
H' 

a substanee which would probably not have acid properties, and which would easily be 
oxidised to dibenzyl phosphinic acid. 

We may mention that quite accidentally we discovered that when the high boiling 
residues obtained in fractionating the primary phosphine are allowed to oxidise spon- 
taneously, the crystalline oil is produced; and as we have shown that the secondary 
phosphine is formed in the sealed tube reaction, we think it quite possible that the 
crystalline oil is its oxide. 

No such body has been obtained hitherto, but then the products of the spontaneous 
oxidation of secondary phosphines have scarcely if at all been investigated. The point is 
one of some interest, and we intend to submit it to enquiry. 

E. " Insoluble Crystalline Body." — We obtained this substance in some of our later 
experiments. It remained undissolved when the oily mass containing the secondary and 
tertiary phosphine was treated with ether, accompanied by oxide of tribenzyl phosphine. 
It was separated from this latter body by boiling with alcohol, in which it was 
insoluble. 

We also found it in the ethereal solution of the phosphines, from which it was slowly 
deposited spontaneously. 

It was insoluble in water, alcohol, and ether, but sparingly soluble in boiling glacial 
acetic acid, from which it was deposited on cooling in colourless needles of characteristic 
crooked form. Twice recrystallised from acetic acid, its corrected melting point was 
found to be 27 6° -27 7°. Two analyses were made, one a combustion with oxide of copper 
and chromate of lead, the other a combustion with pure oxide of copper, in order to 
determine the phosphorus as well as carbon and hydrogen. 

Analysis. 

(l) Combustion with oxide of copper and chromate of lead. 

( 0-6355 C0 2 = 0-173318 C = 75-75 per cent. 
0-^88 gave j . 136() H Q = . 015in H = 6 . 6Q : 



586 PROF. LETTS AND MR R. F. BLAKE ON 

(2) Combustion with pure oxide of copper. 

c 10432 C0 2 =0-284509 C = 7528 per cent. 
03779 gave I 02258 H 2 =0025088 H=663 
( 01277 Mg 2 P 2 7 = 00356639 P= 943 

Obtained. 

Calculated for (CfH^PO., 





I. 


II. 




Carbon, 


75-28 


7575 


7500 


Hydrogen, 


6 63 


660 


6-25 


Phosphorus, . 


9-43 


... 


922 



It is very singular that this substance has the same melting point as sulphide of 
tribenzyl phosphine, and precisely the same solubility relations. The percentage amounts 
of carbon, hydrogen, and phosphorus are also identical. Its crystalline habit is, however, 
different, and we were unable to detect any sulphur in it, whereas by the test we 
employed (ignition with caustic potash and chlorate of potash) we readily found the 
sulphur in a specimen of the sulphide. 

We had not sufficient of the substance to study its properties exhaustively, but we 
ascertained that it readily yielded oxide of tribenzyl phosphine. This we accomplished 
by dissolving it in hot acetic acid, and adding bromine, in excess, when the characteristic 
compound of the oxide separated, which was identified by analysis. 

Analysis. 

0-6789 required 223 c.c^ AgN0 3 = 0-1784 Br = 2627 per cent. 

Mean of ten Determinations of 
Obtained. the Bromine Compound of 

Tribenzyl Phosphine Oxide. 
Bromine, .... 26"27 26-70 

A quantity of the brominated body was boiled with caustic soda solution, and the 
product, after washing, recrystallised. Its melting point was now found to be 215°-215°'5 
— exactly that of the oxide, and it had its crystalline form. 

It is very improbable that a tertiary phosphine peroxide could be formed under the 
conditions in which the substance in question was produced, and as yet no such body 
has been described. But it seems equally improbable, from the properties of the sub- 
stance, that it is the isomer, viz., dibenzyl phosphinate of benzyl, 

C 7 H 7 ^~^-^^ 

C 7 H 7 /P = 0. 

C 7 H 7 -0/ 

We therefore remain in doubt as to its nature. 

We have thus separated from the products of Hofmann's sealed tube reaction the 
following ten bodies (or have proved them to be present) : — 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 587 



(1) Monobenzyl phosphine, .... 

(2) Benzyl phosphinous acid, .... 

(3) Benzyl phosphinic acid, 

(4) Dibenzyl phosphine, ..... 

(5) Dibenzyl phosphine oxide (?), (" crystalline oil "), 

(6) Dibenzyl phosphinic acid, .... 

(7) Tribenzyl phosphine, ..... 

(8) Tribenzyl phosphine oxide, .... 

(9) (''Insoluble crystalline body"), 
(10) A resin containing phosphorus, insoluble in alcohol and ether, but soluble in 

chloroform. 



(C 7 H 7 )H 2 P 

(C 7 H 7 )H 2 P0 2 

(C 7 H 7 )H 2 P0 3 

(C 7 H 7 ) 2 HP 

(C 7 H 7 ) 2 HPO(?) 

(C 7 H 7 ) 2 HP0 2 

(C 7 H 7 ) 3 P 

(C 7 H 7 ) 3 PO 

(C 7 H 7 ) 3 PO 2 0) 



It is not possible for us to say how many of these different substances exist as such in 
the original product of the sealed tube reaction, and how many are formed in the 
subsequent treatment to which it is subjected. 

The three phosphines no doubt exist in the original product as hydriodates or 
hydrochlorates combined with chloride or iodide of zinc, and we think that their produc- 
tion is of great interest, for hitherto only primary and secondary phosphines have been 
obtained when an alkyl iodide, oxide of zinc, and phosphonium iodide are heated together 
under pressure ; whereas we prove that, when chloride of benzyl is substituted for an 
alkyl iodide, the tertiary phosphine is formed also. 

It is well-nigh impossible for us to say definitely whether the quaternary compound is 

formed as well, for it would certainly combine with chloride or iodide of zinc, and would 

be very troublesome to separate from these substances. But the occurrence of oxide of 

tribenzyl phosphine, after the product of the sealed tube reaction had been steamed and 

treated with potash in a hydrogen atmosphere, is, in our opinion, almost conclusive proof 

that tetrabenzyl phosphonium was orignally present, and had been decomposed by the 

alkali, 

(C 7 H 7 ) 4 PI -(- KHO = (C 7 H 7 ) 3 PO + C 7 H 8 + KI , 

a reaction which occurs with the greatest ease when the two substances are boiled to- 
gether in aqueous or alcoholic solution. How otherwise could the oxide be produced in 
quantities so large as those we observed ? 

If we assume that all four phosphorised derivatives are produced, the reaction 
becomes analogous to that which occurs when a haloid compound of a hydrocarbon 
radical is treated with ammonia, the only difference being that hydracid (hydrochloric) 
is liberated in the case of phosphuretted hydrogen, and is found in the gases contained 
in the sealed tubes. 

Whether the substitution of an alkyl chloride for an iodide is the cause of the pro- 
duction of a tertiary and (probably) quaternary compound in addition to the primary 
and secondary phosphines, or whether benzyl behaves differently in Hofmann's reaction 
from other alkyl radicals, we have not as yet decided. There can, however, be very little 



588 PROF. LETTS AND MR R. F. BLAKE ON 

doubt that Hofmann's experiments in the ethyl and methyl series prove conclusively, 
that a tertiary phosphine is not formed when their iodides are employed. 

As regards the other substances isolated from the sealed tube reaction, which consist 
chiefly of the products of the oxidation of the benzyl phosphines, we think it probable 
that benzyl phosphinous acid and benzyl phosphinic acid were produced by the oxida- 
tion of the primary phosphine by the air, for it was impossible to prevent the intro- 
duction of some air while steaming out the contents of the sealed tubes. The same 
remark might apply to the " crystalline oil " and to the dibenzyl phosphinic acid, formed 
in a similar manner from dibenzyl phosphine. But, of course, it is also perfectly 
possible that all three acids and the crystalline oil were produced by some complicated 
interaction with oxide of zinc — and this is very likely to be the case with (some at least 
of) the dibenzyl phosphinic acid, as it was found in the aqueous solution of the product, 
before it had been boiled with potash to liberate the secondary and tertiary phosphine. 

In conclusion, we have only to add that this research has extended over a very long 
period, and has envolved an expenditure of time, energy, and material probably quite 
out of proportion to its value. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 589 

PAET II.— THE ACTION OF ALCOHOLS ON A MIXTURE OF PHOSPHORUS 

AND ITS IODIDE. 

The methods which have as yet been discovered for the preparation of the phosphines 
and their derivatives are, as a rule, very tedious, troublesome, and unsatisfactory in point 
of yield, — a fact which no doubt has considerably retarded the investigation of this 
interesting group of organic compounds, and which soon becomes unpleasantly apparent 
to any one who makes them a study. 

To the best of our belief the following are the only methods of any importance : — 

(1) Action of the Haloid Derivatives of Hydrocarbon Radicals on Metallic Phosphides. 
— It was by this reaction that Paul Thenard # discovered the first members of the 
phosphine group in 1843-1847. He experimented with chloride of methyl and phos- 
phide of calcium, the investigation being attended with considerable difficulties, owing to 
the labour involved in separating the different products and in obtaining them in a pure 
state; also on account of their poisonous properties and their explosive and inflammable 
nature. 

In spite, however, of these difficulties, Thenard apparently isolated trimethyl phos- 
phine, a substance analogous to kakodyle, (CH 3 ) 4 P,, and methylated solid phosphide of 
hydrogen (CH 3 ) 2 P 4 . The latter he described as an inert solid body, the second as a 
spontaneously inflammable liquid, boiling at 250° C. — very explosive, poisonous, and 
unstable. 

Thenard recognised the relationship of trimethyl phosphine to ammonia, and 
predicted the existence of the then undiscovered organic compounds of nitrogen and 
antimony. 

Wurtz in 1848, and Hofmann in 1850, verified Thenard's prediction by discovering 
the compound ammonias ; and Lcewig and Schweitzer a little later obtained stibethyl. 

In 1855 Hofmann and CAHOURst investigated the action of iodide of methyl on 
phosphide of sodium, and obtained, in addition to trimethyl phosphine and the phos- 
phorised kakodyle of Thenard, iodide of tetramethyl phosphonium. They found that 
the reaction was very energetic when iodide of methyl and phosphide of sodium were 
heated together, and that inflammable and explosive substances were formed, rendering 
the method dangerous, and exposing, as they said, the fruits of their labour to loss. 
Moreover, it was unreliable, and furnished mixtures the separation of which presented 
enormous difficulties. 

Berle I about the same time obtained triethyl phosphine by the action of phosphide 
of sodium on iodide of ethyl, the former substance being prepared by heating sodium 
and phosphorus together in rock oil. Iodide of ethyl only acted upon this at a high 
temperature, and only very small quantities of the phosphine were produced. 

* Comptes Rendus, xxi. p. 144, and xxv. p. 892. 

t Hofmann and Cahours, Phil. Trans., 1857 ; and Ann. de Chim. et de Phys. (3), lxi. p. 5. 

| Bekle!, Jour. f. Prac. Chem., lxvi. p. 73. 



590 PROF. LETTS AND MR R. F. BLAKE ON 

Berle next attempted * to prepare the tertiary phosphine by heating sodium, phos- 
phorus, and iodide of ethyl together in a sealed tube, but although a reaction occurred, he 
does not appear to have obtained any very satisfactory results. 

CAHOURst in 1859 prepared iodide of tetrethyl phosphonium by the action of iodide 
of ethyl on crystallised phosphide of zinc (obtained by heating the metal in phosphorus 
vapour at 180° C). 

In 1882 one of us and N. Collie X investigated the action of benzyl chloride on phos- 
phide of sodium. The latter substance was obtained with perfect ease and safety by 
acting upon sodium with phosphorus under the surface of dry xylol. The two bodies 
(chloride of benzyl and the metallic phosphide) react readily when they are heated 
together, and chloride of tetrabenzyl phosphonium is produced in abundance. Experiments 
were also commenced with the haloid derivatives of some other hydrocarbons, but were 
not completed. The results obtained were on the whole satisfactory, and no doubt this 
method for producing tertiary and quaternary phosphorus derivatives will find a more 
extended application in the future. 

(2) Action of Organo-Metallic Bodies on Trichloride of Phosphorus. — The action of 
zinc alkyls on trichloride of phosphorus was first investigated by Hofmann and 
Cahours,§ and was further studied by Hofmann alone. || It is in a sense the 
reverse of the action of a metallic phosphide on a haloid ether, but resembles it in 
that the action in both cases is determined by the attraction of halogen for metal. By 
this method tertiary phosphines are exclusively formed — 

3R 2 Zn + 2PCl 3 =2R 3 P + 3ZnCl 2 . 

It is necessary to treat the product of the reaction with caustic potash, in order to 
decompose the compound of phosphine and chloride of zinc and to liberate the former. 
By its means Hofmann and Cahours obtained trimethyl and triethyl phosphine, and 
submitted them to an exhaustive examination. They showed that tertiary phosphines 
resemble the corresponding amines in many respects, especially in the readiness with 
which they combine with the iodides of hydrocarbon radicals to give quaternary com- 
pounds. On the other hand, they proved that, unlike the amines, tertiary phosphines 
readily combine with oxygen to give very stable compounds of the general formula R 3 PO. 

(3) Action of Alcohols on Phosphonium Iodide; and (4) Action of Alky I Iodides on 
Phosphonium Iodide and Oxide of Zinc (Hofmann's methods). — In the year 1871, 
seventeen years after his experiments with Cahours, Hofmann again took up 
the study of the phosphines, and succeeded in discovering a simple method not only for 
obtaining the tertiary and quaternary compounds, but also the primary and secondary 
bases, — substances which had not been previously obtained, and whose investigation 

* Berle, Comp. Rend., xlix. 

t Cahours, Comp. Rend., xlix. p. 87 ; and Jahresbericht, 1859, p. 430. 

I Letts and Collie, these Transactions, xxx. part 1, p. 181. 

§ Hofmann and Cahours, Ann. de Chirn. et de Phys. (3), li. 

|| Hofmann, Ibid., lxii., lxiii., lxiv. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 591 

led to very interesting and important results. In searching for some general method 
for obtaining the phosphines, Hofmann was influenced by his well-known and beautiful 
researches on the preparation of the compound ammonias. " The question arose," he 
says,* "should not the series of phosphines be capable of production by a general reaction 
similar to that which had yielded me the compound ammonias twenty years ago ? For 
that purpose it was necessary to allow phosphuretted hydrogen to react upon alcohol 
iodides under suitable conditions." 

In his first experiments in this direction, Hofmann heated iodide of phosphonium in a 
sealed tube, in which was also placed a narrower tube containing iodide of ethyl and a 
little water. The sealed tube thus charged was heated in a horizontal position for several 
hours to a temperature ranging from 160°-180° C, when a reaction occurred in the desired 
manner, triethyl phosphine and tetrethyl phosphonium iodide resulting. The yield, how- 
ever, was very small, owing to the large quantity of hydriodic acid set free. To avoid 
this, Hofmann took advantage of the fact that phosphonium iodide when heated with 
alcohol yields phosphuretted hydrogen, iodide of ethyl, and water. Accordingly, he next 
charged the sealed tubes simply with alcohol and phosphonium iodide. When this 
mixture was heated for eight hours at 180° C, the reaction was complete. The cooled 
tubes were found to be full of a snow-white crystalline mass, and on opening them scarcely 
a trace of gas escaped. On adding water, the solid dissolved to a homogeneous liquid, 
showing that no iodide of ethyl remained, neither was any alcohol left undecomposed. 
The crystals showed themselves to consist of a mixture in nearly equal parts, of triethyl 
phosphine hydriodate, and tetrethyl phosphonium iodide. Their separation presented no 
difficulty, for on addition of caustic soda, triethyl phosphine separated as a clear liquid, 
while the solution gave on evaporation beautiful crystals of tetrethyl phosphonium iodide. 

As regards the most favourable proportions of iodide of phosphonium and alcohol 
(which are of great importance), Hofmann found that one molecular weight of the former 
should be taken for three molecular weights of the latter. "In this case," he says, "the 
quantities heated in the tubes can be increased in an extraordinary manner without fear 
of explosions. In one experiment there was placed in a single tube 25 grms. of phos- 
phonium iodide and 22 grms. of alcohol." 

Hofmann next investigated the course of the reaction, to decide whether iodide of 
ethyl was first formed and was then acted upon by the phosphuretted hydrogen, or 
whether the alcohol and phosphonium iodide acted upon each other directly — 

(1) PH 4 I + 3aH a O=(C 2 H 5 ) 3 PHI + 3H 2 

(2) PH 4 I + 4C 2 H 6 0=(C 2 H 5 ) 4 PI +4H 2 0. 

To decide this question the tubes were heated for four hours only. On cooling, two 
layers of liquid were visible. The tubes when opened showed great pressure, and on 
distilling their contents iodide of ethyl passed over in abundance. It may be taken for 
granted then that the reaction occurs in two phases, in the first of which iodide of ethyl 

* Hofmann, Berichte f iv. (1871) 205. 
VOL. XXXV. PART II. (NO. 15). 5 C 



592 PROF. LETTS AND MR R. F. BLAKE ON 

is liberated, which then acts upon phosphuretted hydrogen in the same way as it acts 
upon ammonia. Hofmann had thus discovered a simple and elegant general reaction 
for the preparation of tertiary and quaternary compounds, which he employed successfully 
in the methyl,* ethyl,t propyl, J butyl,§ and amyl|| series. 

But the primary and secondary bases were still wanting, though the possibility of 
their existence could scarcely be doubted in view of the well-established analogies exist- 
ing between the derivatives of ammonia and phosphuretted hydrogen. Hofmann there- 
fore again took up the matter, and began a new series of investigations which led to 
brilliant results. 

The direction of these new experiments was, as he says, clearly indicated, for as both 
the tertiary and quaternary derivatives are obtained by direct substitution from phos- 
phuretted hydrogen, so also ought the primary and secondary bases to be produced. For 
if three and four molecules respectively of the alcohol can be made to react upon one 
molecule of phosphonium iodide, one and two molecules ought also to react in a similar 
manner under suitable conditions — 

(1) PH 4 I + C 2 H 6 =(C 2 H 5 )PH 2 .HI + H 2 0. 

(2) PH 4 I + 2C 2 H 6 = (C 2 H 5 ) 2 PH.HI + 2H 2 . 

Hofmann IT therefore sought to achieve the desired object by altering the proportions 
of phosphonium iodide and alcohol, but without success, the tertiary base being alone 
produced, or a mixture of the tertiary and quaternary compounds ; while, with the 
proportions required for the second of the above equations, the tubes invariably exploded. 

In the meantime Drechsel and Finkenstein ## believed that they had succeeded in 
obtaining the primary bases of the ethyl and methyl series by saturating the iodides of 
those radicals with phosphuretted hydrogen, and allowing the solutions to remain for 
some time at ordinary temperatures or by heating them to 100° C. Under these 
conditions a crystalline body was obtained, which was not phosphonium iodide, as it 
dissolved in water without evolution of gas, and which they therefore concluded to be 
the hydriodate of the primary base. 

Also, by heating an ethereal solution of iodide of zinc saturated with phosphuretted 
hydrogen, with iodide of methyl, they believed that they had obtained the same 
substance. 

Hofmann tt repeated these experiments, and showed that the primary bases were not 
formed at all, whereas the tertiary and quaternary derivatives were. 

The idea then occurred to him of heating the alkyl iodide with iodide of phosphonium 
in presence of a metallic oxide, with the happiest results ; for on heating a mixture of 
phosphonium iodide, ethyl iodide, and oxide of zinc, in the proportions of two molecules 
of each of the former to one molecule of the latter in sealed tubes at 150°C. for from six to 

* Hofmann, Berichte, iv. (1871) p. 209. t Hofmann, Berichte, iv. (1871) p. 205. 

\ Ibid., Berichte,, vi. (1873) p. 292. § Ibid., Berichte, vi. (1873) p. 296. 

|| Ibid., Berichte, vi. (1873) p. 297. 1 Ibid., Berichte, iv. (1871) p. 372. 

** Drechsel and Finkenstein, Berichte, iv. (1871) p. 352. tt Hofmann Berichte, iv. (1871) p. 372. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 593 

eight hours, a complete reaction occurred, the tubes when cold containing a crystalline mass 
consisting exclusively of the hydriodates of the primary and secondary bases, the former 
being the chief product, while the latter was only formed in small quantities. The 
separation of the two was accomplished in the simplest manner possible, the addition of 
water to the product of reaction liberating the primary phosphine alone, which was dis- 
tilled off in a hydrogen atmosphere, while the secondary phosphine was subsequently set 
at liberty by the action of an alkali. 

Thus Hofmann discovered for the first time a general method for preparing primary 
and secondary phosphines, the reaction being exactly complementary to that by which 
he had obtained the tertiary and quaternary compounds ; like it the method was 
distinguished by its simplicity and the readiness with which it may be carried out. 

By its means he prepared methyl,* ethy],t propyl,^ butyl, § amyl,|| and benzyl 1 
phosphines. 

(5) Michaelis' Methods. — Hofmann's methods, although of excellent service for 
obtaining the phosphines of those radicals which form alcohols, could not be employed 
in the preparation of phosphines containing purely aromatic radicals. He in vain 
endeavoured to obtain phenyl phosphines, ## but neither by the action of chlorobenzol 
on phosphonium iodide, nor by heating a mixture of phenol and phosphonium iodide, 
was a trace of those substances produced, and he was equally unsuccessful in his 
attempts to prepare tolyl phosphines. 

Michaelis, on the other hand, by approaching the question from an entirely different 
direction, not only succeeded in obtaining all the phenyl phosphines, but also in discover- 
ing a fairly general method for the production of primary phosphines. The substance 
forming the starting-point for the preparation of phosphines by these methods is trichloride 
of phosphorus. One atom of chlorine is first replaced in that body by one or other of 
the following processes :- — 

(1) The mixed vapours of a hydrocarbon and the trichloride are repeatedly passed 

through a red hot tube. Thus, when benzol is employed, ''phosphenyl" chloride is 

obtained — 

PC1 3 +C 6 H 6 = (C 6 H 5 )PC1 2 +HCL 

(2) A mercury ether is heated with the trichloride under pressure. Thus, when 
mercury ethyl is employed, ethyl-phosphorus chloride is obtained together with ethyl 

chloride of mercury — 

PCl 3 -l-(C 2 H 5 ) 2 Hg = (C 2 H 5 )I > Cl 2 +Hg(C 2 H 5 )Cl. 

(3) By digesting a hydrocarbon with the trichloride and aluminum chloride, the 
reaction being the same as 1. Other substances besides hydrocarbons yield substituted 
phosphorus chlorides when submitted to this reaction. Thus a mixture of acetone, 

* Hofmann, Berichte, iv. (1871) p. 430. t Hofmann, Ibid., iv. (1871) p. 605. 

X Ibid., vi. (1873). p. 292. § Ibid., vi. (1873) p. 296. 

|| Ibid., vi. (1873) p. 297. II Ibid., v. (1872) p. 100. 
** Ibid., Berichte, v. (1872) p. 100. 



594 PROF. LETTS AND MR R. F. BLAKE ON 

phosphorus trichloride, and aluminum chloride react spontaneously, the reaction 
proceeding according to the equation — 

2(CH 3 ) 2 CO + PC1 3 = 2HC1 + (CH 3 - CO - CH 2 ) 2 PC1 . 

By means of these different reactions Michaelis and his pupils have obtained a 
considerable number of substituted phosphorus-chlorides, among which are — 

Phenyl * phosphorus-chloride by methods 1, 2, and 3. 
Tolylf „ „ 3 



XylyU 

Ethyl § 
Propyl § 
Naphthyl || 
Ace tony 1 1T 



The substituted chlorides resemble trichloride of phosphorus itself, not only in 

composition, but also in properties. As a rule they are fuming liquids, combining readily 

with chlorine to give solid compounds analogous to pentachloride of phosphorus ; treated 

with water they yield phosphinous acids. Thus phenyl phosphorus chloride gives phenyl 

phosphinous acid — 

(C 6 H 5 )PC1 2 +2H 2 0=(C 6 H 5 )PH 2 2 + 2HC1. 

While the products of addition which they form with chlorine react with water to 

give phosphinic acids — 

(C 6 H 5 )PC1 4 + 3H 2 0=(C 6 H 5 )PH 2 3 +4HC1. 

Having obtained the substituted chlorides, Michaelis sought for some method for con- 
verting them into the corresponding primary phosphines, but the reduction proved to be 
more difficult than he anticipated. In his first experiments (with phenyl phosphorus 
chloride) he endeavoured to accomplish it by means of nascent hydrogen (evolved by the 
action of hydrochloric acid on zinc dust, and also by that of glacial acetic acid on zinc or 
on sodium amalgam). In these experiments only a trifling reduction occurred, mere 
traces of the primary phosphine being formed. 

He next studied the action of hydriodic acid on the chloride, and found that when 
the gaseous hydracid was passed into the chloride, chlorine was gradually replaced by 
iodine and the resulting iodide combined with hydriodic acid, thus — 

C 6 H 5 PC1 2 + 3HI = C 6 H 5 PI 2 HI + 2HC1 . 

From this iodide Michaelis obtained phenyl phosphine by a similar reaction to that 
which gives rise to phosphuretted hydrogen from iodide of phosphorus, only instead of 
employing water for its decomposition alcohol was used. He gives the following equation 
as representing the reaction which occurs — 

3C 6 H 5 PI 2 ,HI + 9C 2 H 6 = C 6 H 5 PH 2 + 2C 6 H 5 P0 3 H 2 + 3H 2 + 9C 2 H 5 I . 

In later experiments he abandoned this method and employed a much simpler one, 

* Berichte, vi. (1873) 601 ; viii. (1875) 922 ; xii. (1879) 1009. t Berichte, xiii. (1880) 653. 

X Annalen, 212, pp. 203 and 209. § Ibid., xiii. (1880) 2174. 

|| Berichte, ix. (1876) 1051. H Ibid., xvii. (1884) 1273. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 595 

namely, the destructive distillation of the phosphinous acid (which, as already stated, is pro- 
duced by the action of water or alcohol on the substituted chloride # ). The phosphinous 
acid decomposes in an analogous manner to phosphorous acid when heated, the products 
of the reaction being the corresponding phosphinic acid and the primary phosphine — 

3C 6 H 5 P0 2 H 2 =C 6 H 5 PH 2 +2C 6 H 5 P0 3 H 2 . 

The following details illustrate the application of this method to the preparation of 
phenyl phosphine : — 

100 grms. of the crude phenyl phosphorus chloride are gradually mixed with excess 
of alcohol and the mixture filtered. The excess of alcohol, &c, is then distilled off in a 
current of carbonic anhydride and the syrupy residue placed in a distilling flask (through 
which a current of carbonic anhydride is passed), and heated over the naked flame. At 
first a little alcohol distils, then at 250° phenyl phosphine passes over. The flame may 
be removed when once the reaction has commenced, as it continues by itself. 14 grms. 
of the pure phosphine, or 60 per cent, of the theoretical yield, may be thus obtained. 

From phenyl phosphorus chloride Michaelis obtained diphenyl phosphine by the 
following reactions : — 

(1) The chloride is digested with mercury diphenyl at 200°, when diphenyl phos- 
phorus chloride results t — 

(C 6 H 5 )PCl 2 +Hg(C 6 H 5 ) 2 = (C 6 H 5 ) 2 PCl+HgCl(C 6 H 5 ); 

or it is heated for some time at a temperature of 280° C, when the following reaction 

occurs — 

2(C 6 H 5 )PCl 2 = (C fi H 5 ) 2 PCl+PCl 3 . 

(2) Diphenyl phosphorus chloride, when heated with water or dilute soda solutioD, 
decomposes in the following manner : J — 

2(C 6 H 5 ) 2 PCl-f2H 2 = (C 6 H 5 ) 2 PH + (C 6 H 5 ) 2 HP0 2 +2HCl. 
He also obtained the tertiary base : at first by acting upon a mixture of phenyl phos- 
phorus chloride and bromide of phenyl with sodium § — 

(C 6 H 5 )PC1 2 + 2(C 6 H 5 )Br + 3Na = 2NaCl + NaBr + ( C 6 H 5 ) 3 P ; 
but later this method was modified in a remarkable manner by substituting for phenyl 
phosphorus chloride, phosphorus chloride alone, the reaction occurring quite easily and 
very energetically at ordinary temperatures, according to the equation — 

PCl 3 +3(C 6 H 5 )Br+6Na = 3NaCl + 3NaBr+(C 6 H 5 ) 3 P. 

* By modifying the conditions of the experiment a totally different reaction may be made to occur. The 
phosphinous acid is obtained by adding the chloride to excess of water ; whereas by employing an insufficient quantity, 
there are produced in addition phenyl phosphinic acid, diphenyl phosphinic acid, and a solid compound of phenyl, 
hydrogen, and phosphorus, (C 6 H S )HP 4 (phenylated solid phosphide of hydrogen). Michaelis gives the following 
equations : — 

(1) (C 6 H 6 )PC1 2 +2H 2 0=(C 6 H 5 )PH 2 2 +2HC1. 

(2) (C 6 H 5 )PC1 2 +(C 6 H 5 )PH 2 2 = 2(C 6 H 6 )PO + 2HC1 . 

(3) 5(C 6 H 5 )PO=(C 6 H 4 ) 4 P 2 3 + P 2 +C 6 H 5 P0 2 . 

(4) 5(C 6 H 5 )PO+H 2 0+3P 2 =2(C 6 H 5 )P 4 H+3C 6 H 5 P0 2 = 2(C 6 H 6 ) 2 P 6 2 H + C 6 H 5 P0 2 . 

t Berichte, viii. 1304. J Michaelis and Gleichmann, Berichte, xv. (1882) p. 801. 

§ Michaelis and Gleichmann, Ibid., p. 820. 



596 PROF. LETTS AND MR R. F. BLAKE ON 

A perusal of the preceding sketch of the different methods hitherto discovered for 
the preparation of the phosphines will be sufficient to convince any one that to obtain 
these bodies in quantity is no easy task, for in nearly every case operations are 
involved which demand considerable expenditure of time and material, and frequently 
the yield of product is small. 

One of us had for a long time sought for some easier and more direct method than 
any of those mentioned, and had made many attempts with that object in view in 
different directions, but without success. 

At last the idea presented itself that the germ of a new process for obtaining phosphines 
ought to exist in the well-known and ordinary method for preparing phosphonium iodide. 

In this reaction, as every one knows, water is allowed to act upon a mixture of 
phosphorus and its iodide, when iodide of phosphonium is obtained on the one hand, 
while phosphoric and phosphorous acids are produced on the other. 

Now, it seemed not unreasonable to suppose that analogous reactions would occur, if 
instead of water an alcohol were used, and if this were the case a new and simple process 
for obtaining phosphines would be at hand. 

Our first experiments to test the truth of this surmise have been made with benzyl 
alcohol, and the results have been unexpectedly gratifying, for not only does the reaction 
occur with the greatest ease, but the number of phosphorised derivatives of benzyl 
produced is remarkably large. Moreover, we have without the slightest difficulty obtained 
most of them in considerable quantity, so that we have been able to make a fairly 
exhaustive examination of their properties. 

In short, the method has placed at our disposal a number of highly interesting 
substances, which could not have been obtained in sufficient quantity by any other means 
as yet known, and we have thus been enabled to fill up many lacunae in the history of 
the benzyl phosphines and their derivatives. We may, however, state at the outset that 
the reaction which occurs between benzyl alcohol, iodide of phosphorus, and phosphorus, 
does not take place in exactly the manner we anticipated. 

TJie Action of Benzyl Alcohol on a Mixture- of Phosphorus and its Iodide. 

We may remark at the outset that we employed in all our experiments the same 
mixture of phosphorus and phosphorus iodide as is used for the preparation of phos- 
phonium iodide. It was prepared as follows : — 40 parts by weight of vitreous phosphorus 
(carefully dried) were dissolved in an equal weight of dry and freshly-redistilled 
bisulphide of carbon, and to the solution 68 parts of iodine were gradually added. The 
retort in which the mixture was made was then heated in a water-bath and a current 
of dry carbonic anhydride passed through it until every trace of bisulphide had distilled 
off. We shall call the product, for the sake of brevity, " the phosphorus mixture." 

We may also mention that in most of our experiments the quantity of benzyl alcohol 
taken was about the equivalent of the quantity of water used in the preparation of 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 597 

phosplionium iodide. Thus, for the above quantities of phosphorus and iodine 24 parts 
of water are required, and the equivalent quantity of benzyl alcohol (H 2 : C 7 H 8 0) is 
144 parts. 

Experiment 1. — 30 grms. of the phosphorus mixture were placed in a flask fitted 
with an upright condenser, and 33 grms. of benzyl alcohol were added. 

The mixture became warm on shaking, but no very energetic reaction occurred until 
the flask containing it was heated in a paraffin-bath ; then, however, an action began and 
increased so rapidly as to become explosive, — the mixture being shot out of the con- 
denser, while dense white vapours were evolved, evidently of iodide of benzyl, from their 
intolerably irritating effect on the eyes. 

A second experiment, conducted in the same way and with the same quantities, led to 
precisely similar results. 

What was left of the product from these two experiments consisted of a brownish 
resinous mass. It was first extracted with ether, which, however, did not dissolve very 
much, then the residue was boiled with alcohol and a solution obtained which deposited 
abundance of crystals (A) on cooling. These were filtered off and the mother-liquors 
concentrated, when a second crystalline product (B) was obtained. 

(A) was recrystallised from spirit. It contained iodine, and when dried became very 
brown. A determination of iodine gave the following result : — 



0-4863 gave 00488 Agl = 00263 1 = 54 per cent. 



It was evidently an impure product, so it was again dissolved in hot spirit. As the 
solution cooled, two distinct sets of crystals separated. One had the colour of bichromate 
of potash, while the other was colourless. 

Eventually it was found that the red crystals were much less soluble in spirit than 
the colourless ones, and by repeated recrystallisation the latter were separated and 
obtained of the constant melting point 192° C. (corr.), while (B), when purified, had the 
same melting point. 

The appearance and melting point led us to believe that the colourless crystals from 
(A) and (B) were both dibenzyl phosphinic acid, and this belief was strengthened by the 
fact that they dissolved readily in a solution of caustic baryta. The solution thus 
obtained was treated with carbonic anhydride (to remove excess of baryta), filtered and 
evaporated to small bulk. As the solution cooled, colourless plates separated out, which 
on analysis gave numbers showing that they were by no means pure. 

Analysis. 

0-4156 gave 0-106 BaS0 4 = -06256 Ba = 15'05 per cent. 
0-3930 lost at 110° C. 00893 H,0 = 2270 „ 

Barium, ..... 
Water, 

All the impure solid products were united and boiled with solution of caustic baryta 



Obtained. 


Calculated for {(C r H r ) 2 P0 2 } 2 Ba,8H 2 


1505 


17-75 


22-70 


1867 



598 PROF. LETTS AND MR R. F. BLAKE ON 

to remove dibenzyl phosphinic acid. The residue was washed with water and dilute 
hydrochloric acid, and then boiled with alcohol. The filtered solution gave on cooling a 
crop of colourless needles, mixed with a few of the red crystals, but by recrystallisation 
they were obtained pure. They had a melting point of 212° C. (uncorr.), agreeing with 
that of tribenzyl phosphine oxide, and they also gave the compound with iodide of zinc, 
microscopically identical with a specimen prepared with a known sample of the oxide. 
The red crystals we did not obtain in sufficient quantity to examine. 

Experiment 2. — In order to avoid the very violent action which had occurred in our 
first experiments, we now allowed the benzyl alcohol to drip down the inverted 
condenser slowly on to the phosphorus mixture from a tap funnel. The first drops 
occasioned a very violent action, even in the cold, the flask becoming full of white fumes 
and iodine vapour ; but as more of the benzyl alcohol was added the action moderated, 
though it continued to be pretty brisk, and sufficiently so to keep the mixture in 
ebullition. When all the alcohol had been thus cautiously added, and the reaction 
appeared to be over, a little more of the phosphorus mixture was added, and further 
action occurred, but it was not very violent. Finally, the flask containing the product 
of the reaction was heated for some time in an oil-bath and then allowed to cool. 

This product consisted of a dark brown viscid mass. To it a quantity of water 
was added, when heat was developed. The flask in which it was contained was 
next connected with a Liebig's condenser, and a current of steam blown through it, when 
a somewhat brisk action occurred, and an oily liquid distilled over. This was in fairly 
large quantity, and was readily identified by its boiling point (111° C.) and odour as 
toluol. The residue in the flask after steaming, consisting of a dark brown oily substance 
swimming on the surface of the water, rapidly solidified. The aqueous solution was 
decanted, and the residue (which had by this time completely solidified) was extracted 
two or three times with boiling water. The aqueous solutions were mixed, and gave on 
cooling an abundant colourless crystalline precipitate (A). 

The residue after thorough extraction with water was pounded up in a dish and 
extracted several times with baryta solution, then washed with water, and finally boiled 
several times with alcohol, until nothing remained apparently but amorphous phosphorus. 
The alcoholic solutions were evaporated down, and gave on cooling an abundant crop of 
colourless needles (B). 

The mother-liquors from (A) were pressed out through linen and evaporated until 
fumes of hydriodic acid came off. They were then allowed to cool, when the solution 
solidified to a radiating crystalline mass (C). The mother-liquors were squeezed out 
through linen and heated on a steam-bath for a considerable time, during which they 
not only fumed, owing to evolution of hydriodic acid, but also effervesced, and the 
effervescence increased on stirring. On heating some of the liquid in a tube it frothed 
considerably, and iodide of phosphonium sublimed, while later the odour of monobenzyl 
phosphine became very distinct and its hydriodate volatilised. 

Examination of Product A. — This was freed as far as possible from adhering mother- 



BENZYL PHOSPHTNES AND THEIR DERIVATIVES. 599 

liquors and warmed with baryta solution, when most of it dissolved. The solution was 
mixed with the baryta extracts of the insoluble product (from which B was eventually 
obtained), and after removing the excess of baryta by a current of carbonic anhydride, 
the filtered solution was evaporated down, once more filtered from a white insoluble 
substance which had separated, and then allowed to cool, when a crop of crystals was 
deposited, consisting of thin plates. These were rapidly washed and dried, first in air 
and then at 110° C. 

Analysis. 

(1) 21655 gave 07852 BaS0 4 = 046168 Ba = 21-31 per cent. 

(2) 1-6808 „ 06116 „ =035961 „ =2139 

As the salt dissolved in hot alcohol and readily separated as the solution cooled, a 
quantity was thus recrystallised, the crystals dried as before, and a barium determination 

made — 

(3) 1-6086 gave 05918 BaS0 4 = 0-34796 Ba = 2163 per cent. 

The numbers obtained show that the salt was dibenzyl phosphinate of barium. 

Obta ined. Calculated for {(C 7 H 7 ) 2 P0 2 } 2 Ba 



I. II. III. 

Barium, . . 2131 2139 2163 2184 

On adding hydrochloric acid to its solution dibenzyl phosphinic acid was precipitated 
in the characteristic form. 

Examination of Product B. — This product consisted of colourless needles. They 
were thoroughly drained in a filter, washed well with spirit, and a melting-point 
determination made, which agreed fairly well with that of tribenzyl phosphine oxide. 
Unfortunately the pure product was accidentally lost, so that no combustion of it was 
made. For that purpose a product obtained from the mother-liquors was used, which 
was once recrystallised. 

Analysis. 

MW(! f 10329 C0 2 = 02817 C =799 per cent. 

3526 grms. gave < „ * r 

s 5 I 0-2262 H,0 = 0025133 H= 7-12 „ 



Obtained. 


Calculated for (C 7 H 7 ) 3 PO 


79-90 


78-75 


712 


656 



Carbon, .... 
Hydrogen, .... 

Although both the carbon and hydrogen were too high, there could be no question 
regarding the identity of the substance, for it gave all the reactions of tribenzyl 
phosphine oxide, especially the very characteristic yellow crystalline compound with 
bromine. Moreover, in other experiments a number of these compounds was obtained 
and analysed. 

Examination of Product C. — The crystals were colourless, and consisted of tufts of 
radiating needles. They were readily soluble in water and in alcohol. Their aqueous 
solution reacted as follows : — With, 

(l) Acetate of lead, an immediate white precipitate. 

VOL. XXXV. PART II. (NO. 15). 5 D 



600 PROF. LETTS AND MR R. F. BLAKE ON 

(2) Sulphate of zinc, no precipitate. 

(3) Sulphate of magnesium, no precipitate. 

(4) Caustic baryta, no precipitate ; but on warming the neutral solution a crystalline 

salt at once separated. 

This behaviour with baryta seemed to afford a method for separating the product 
from phosphoric and hydriodic acids, and for obtaining it in a state of purity. 

Accordingly, the whole of it (after removal of adhering mother-liquor by pressure 
between filter-paper) was dissolved in warm water, and bartya added until the solution 
was nearly neutralised. A slight white precipitate was filtered off, and the filtrate heated 
gradually in a water-bath, when a considerable quantity of a salt separated in very thin 
iridescent plates, which were collected on a filter, washed and dried at 110° C. 

Analysis. 

1-6607 gave 1-2748 BaS0 4 = 074956 Ba = 4513 per cent. 

Obtained. Calculated for C 7 H 7 P0 3 Ba 

Barium, 4513 44-62 

The mother-liquors from which this salt had separated, when evaporated to small 
volume, gave a very soluble crystalline salt, which was also analysed. 

Analysis. 

1-4198 lost at 110° C. 01457 H 2 = 10-26 per cent. 

1-4198 gave 0-6232 BaS0 4 = 0-36643 Ba = 2579 

Obtained. Calculated for (C 7 H 7 HP0 3 ) 2 Ba,3H 2 

Water, 1026 1013 

Barium, 25-79 25-61 

There could then be very little doubt as to the composition of the original product. 
It was benzyl phosphinic acid and its two barium compounds which were analysed, the 
normal and acid salts respectively. To be quite certain on this point, a quantity of the 
normal barium salt was decomposed with the exact quantity of sulphuric acid, and the 
filtered solution evaporated down. The acid then separated in colourless crystalline 
crusts. These were dried by pressure and recrystallised from water, then dried, first on 
filter-paper and subsequently at 110° C. 

Analysis. 

( 06485 C0 2 = 0-176863 C =48-70 per cent. 
0-3631 gave j Q . 1842 HgQ = . 0204666 H = 5-63 „ 

Obtained. Calculated for (C 7 H 7 )P0 3 H 2 

Carbon, 48-70 4883 

Hydrogen, 5"63 5"23 

Monobenzyl phosphinic acid had not been obtained previously. We describe its 
properties and salts on p. 612. 

In the experiment, the results of which we have just described, we thus obtained and 
analysed three different phosphorised benzyl derivatives, which may be considered as the 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 601 

products of the oxidation of the primary, secondary, and tertiary phosphines respectively ; 
while, on concentrating the mother-liquors from the monobenzyl phosphinic acid, we 
obtained a fourth, viz., monobenzyl phosphine, which we suspected to originate from a 
fifth, namely, benzyl phosphinous acid, C 7 H 7 .P0 2 H. 2 . 

Fresh experiments were therefore necessary, not only to decide whether this latter 
compound is really produced, but also to ascertain the quantities of the different 
substances formed, the nature of the reactions which yield them, and the conditions under 
which they are originated. 

Experiment 3.— Three quantities, each consisting of 50 grms. benzyl alcohol and 50 
grms. of the phosphorus mixture (very carefully prepared with pure materials and 
thoroughly free from bisulphide of carbon), were worked up as follows : — 

The weighed quantity of phosphorus mixture was placed in a flask of about 500 c.c. 
capacity. The flask was then filled with carbonic anhydride, attached to an upright 
condenser, and the benzyl alcohol allowed to drop from a tap funnel down the condenser. 
Action occurred spontaneously as soon as some of the phosphorus mixture became 
moistened with the alcohol, and at first violet vapours of free iodine appeared. It 
seemed better to run in the alcohol in tolerably large quantities, so as to keep up a brisk 
action, sufficient to cause rapid ebullition, but care had to be exercised, or the reaction 
became unmanageable. When most of the alcohol had been added the reaction grew 
sluggish, but was started afresh by shaking the flask, so as to thoroughly mix its 
contents. 

In the case of two of the quantities the flasks were heated after all action was over, 
but we think this was not advisable, as a sudden puff of hydriodic acid occurred. Always 
towards the close of the reaction a little phosphonium iodide sublimed in the 
condenser. 

The product in each case consisted of a dark brown resinous mass. About 300 c.c. 
of cold water were next added to the contents of each flask and thoroughly incorporated 
with them by shaking, when the mixture grew slightly warm. A current of steam was 
next passed as long as oily liquid distilled over. By this means, from the three 
quantities operated upon, 45 grms. of toluol were obtained boiling at 110°-i20°C. A 
small quantity of a less volatile liquid remained above this temperature, but its amount 
was trifling. Each of the products was steamed with three separate quantities of water, 
so as to extract all soluble substances as thoroughly as possible. The residue was well 
squeezed in a linen bag, and repeatedly digested with baryta solution as long as anything 
was dissolved. For this purpose the mass was pounded in a mortar with hot baryta 
solution, the extract decanted through a linen filter, and the operation repeated. Finally, 
the residue was well squeezed in a linen filter, and next again and again boiled with 
alcohol so long as anything was dissolved out. What remained undissolved consisted 
apparently of amorphous phosphorus, and weighed 35 grms. 

The product of the reaction was thus split up into — 

(1) Aqueous extracts, containing hydriodic acid, monobenzyl phosphinic acid, some 



602 PROF. LETTS AND MR R. F. BLAKE ON 

dibenzyl phosphinic acid (which rapidly separated from the solution almost completely), 
and other soluble substances. 

(2) Baryta extracts, containing chiefly dibenzyl phosphinate of barium. 

(3) Alcoholic extracts, containing tribenzyl phosphine oxide. 

(4) Insoluble matter, chiefly amorphous phosphorus. 
The treatment of the different extracts was as follows : — 

Aqueous Extracts. — These were filtered from the dibenzyl phosphinic acid which had 
separated out,* then evaporated to small bulk, and allowed to crystallise, when a solid 
mass of benzyl phosphinic acid resulted. This, after squeezing as far as possible in a 
linen bag, weighed 28 grms. The mother-liquors were evaporated in a water-bath until 
dense fumes of hydriodic acid came off, and they were then allowed to cool, when another 
quantity of the phosphinic acid was obtained, which weighed, after drying by pressure, 
32 grms. The mother-liquors, on further concentration, gave no more phosphinic 
acid. Thus, in all nearly 60 grms. of crude acid were obtained. 

A quantity of the mother-liquors when heated in a distilling flask gave at first a 
distillate of hydriodic acid of constant boiling point, then phosphuretted hydrogen (some 
of which exploded in the condenser), a little iodide of phosphonium, and later hydriodate 
of monobenzyl phosphine. Another portion of the mother-liquors was heated in a 
distilling flask immersed in an oil-bath, and lost about half its volume of hydriodic 
acid of constant boiling point. The viscous residue was diluted with water, and 
neutralised with baryta, when a large quantity of a white amorphous precipitate was 
formed. It was filtered off, washed and analysed, and was found to consist of barium 
phosphate. 

Analysis. 

11814 lost when ignited 0-0460 = 389 per cent. 

2-2540 (dried at 110° C.) gave 2-2634 BaSO 4 = T33084 Ba=5904 per cent. 

Obtained. Calculated for BaHP0 4 

Barium, ...... 59"04 58'80 

Water, 3-89 3-86 

30 grms. of barium phosphate were obtained from the portion of the mother-liquors 
experimented upon, and it was estimated that 50 grms. would have been obtained if the 
whole of the mother-liquors had been employed. 

Baryta Extracts. — These contained dibenzyl phosphinic acid almost exclusively. 
They were evaporated to dryness, and the residue heated for a day at 110° C. It weighed 
68 grms., which correspond with 53*3 grms. of the acid. It was redissolved in water, and 
the acid obtained by precipitation with sulphuric acid and extraction of the precipitate 
with alcohol. 

Alcoholic Extracts. — These contained principally tribenzyl phosphine oxide. They were 
evaporated to small bulk and allowed to cool, when the crude oxide separated. It 

* This was well washed, dissolved in baryta, and the solution added to the baryta extracts. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 603 

weighed, when dried as far as possible by pressure, 30 grms. The residual mother- 
liquors left on evaporation 3 to 4 grms. of a viscous mass, which was not further 
examined. 

Experiment 4. — We had as yet not proved the production of benzyl phosphinous 
acid, though we had reason for suspecting its presence, as well as that of either 
phosphorous or hypophosphorous acids. For, on heating the mother-liquors from the 
crude phosphinic acid, we had noticed the production of the primary phosphine (or rather 
its hydriodate) and phosphonium iodide, and their formation seemed to be most readily 
explained on the hypothesis that the acids we have named were present originally, 
but were decomposed by heat. 

To decide this question we performed a new experiment.* 

Three quantities of benzyl alcohol, amounting in all to 145 grms., and three quantities 
of the phosphorus mixture, amounting to 200 grms., were worked up as described in 
Experiment 3, with this difference, that the benzyl alcohol was run in rather more slowly, 
so that the action was not quite so violent. The treatment of the crude product was at 
first substantially the same as in the previous experiment. 

The tribenzyl phosphine oxide, however, at once showed itself to be impure, for 
when dried it became brown. It would not dissolve completely in cold acetic acid, 
whereas the pure oxide dissolves with the greatest ease. On boiling with acetic acid 
it entirely dissolved, but as the solution cooled, a substance separated in feathery crystals 
like sal-ammoniac, quite unlike the oxide of tribenzyl phosphine, which crystallises in 
needles. From the appearance of these crystals we were led to suspect that they were 
iodide of tetrabenzyl phosphonium, — a suspicion which was verified both by their analysis 
(after recrystallisation from alcohol) and by their properties. 

Analysis. 

0-7732 gave 03390 AgI=0\L83204 1=23-69 per cent. 

Obtained. Calculated for (C 7 H 7 ) 4 PI 

Iodine, 23-69 24-32 

When boiled with alcoholic soda solution they were decomposed, and on washing with 
water and recrystallising from alcohol, needles were obtained, melting at 216° C. (corr.). 
(The pure oxide melting at 215°-5.)t Some of the crystals were dissolved in alcohol, 
decomposed with oxide of silver, the product boiled with hydrochloric acid, and extracted 
with boiling water. The solution gave on cooling the characteristic needles of chloride 
of tetrabenzyl phosphonium, a further proof that the original substance was the iodide. 

The main object of this experiment was, as we have said, to ascertain whether benzyl 
phosphinous acid was formed in the reaction. As it is a very soluble substance, we 
expected to find it, if it was produced at all, in the mother-liquors from the phosphinic 

* In the meantime we had obtained benzyl phosphinous acid from other sources, and had become acquainted with 
some of its properties. Among them, we had ascertained that it yields the primary phosphine when heated. 

t One of us and N. Collie have shown that the haloid salts of tetrabenzyl phosphonium are decomposed when 
treated with an alkali, and yield the tertiary phosphine oxide (see p. 575). 



604 PROF. LETTS AND MR R. F. BLAKE ON 

acid, and we had already had some evidence of its presence there. We searched for it 
by the following method : — 

The mother-liquors remaining, after the phosphinic acid had been separated as far 
as possible by concentration and crystallisation, were diluted and mixed with excess of 
acetate of lead. The insoluble lead salts produced (chiefly iodide and phosphate) were 
filtered off and the excess of lead removed by a current of sulphuretted hydrogen. The 
solution was filtered and evaporated until all acetic acid was driven off, and nothing but a 
syrupy liquid remained, which gave off benzyl phosphine on heating. The investigation 
of this syrup, which evidently contained the phosphinous acid, proved troublesome. A 
portion was neutralised with chalk and evaporated to small bulk, when a salt separated 
in crystalline crusts. 

Analysis. 

0-3965 (dried at 110° C.) gave 0-0768 CaO=0-0548 Ca = 13-82 per cent. 

Obtained. Calculated for { (C 7 H 7 )HP0 2 },,Ca 

Calcium, . . . . ■ . 13-82 11-42 

This showed that the calcium salt was not pure. 

The rest of the syrup was neutralised with baryta and evaporated to small bulk, 
when the solution crystallised on cooling, and apparently two sets of crystals were formed. 
They were dried as far as possible by pressure, dissolved in a very little water, and then 
alcohol was added to the solution, when a crystalline salt was precipitated. This was 
washed and analysed. 

Analysis. 

06294 (dried at 110° C.) gave 05456 BaSO 4 = 0-3208 Ba = 5096 per cent. 

Obtained. Calculated for (PH 2 2 ) 2 Ba 

Barium, 50-96 51-31 

The salt gave all the reactions of a hypophosphite. Heated, it evolved phos- 
phuretted hydrogen ; boiled with sulphate of copper, the copper was reduced, &c. 

The mother-liquors from which it had separated gave on evaporation in the desiccator 

a cake of crystals. Some of these were drained on filter-paper, then dried at 110° C. until 

they were of constant weight. 

Analysis. 

0-5564 gave 0-3251 BaSO 4 =01911 Ba=34-34 per cent. 

Obtained. Calculated for {(C 7 H 7 )P0 3 H} 2 Ba 

Barium, 3434 30-64 

A solution of this salt behaved with acetate of copper in the manner so characteristic 
of benzyl phosphinite, giving a greenish precipitate in a strong solution, no precipitate 
in a weak solution, but an immediate green precipitate on boiling. Moreover, the mother- 
liquors from tlie copper salt obtained from a strong solution deposited the green salt 
immediately on boiling. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 605 

As the percentages both of calcium and barium in the respective salts analysed 
indicated that they did not consist of the pure phosphinites, but contained some hypo- 
phosphite, it was determined to obtain the zinc salt, which is deposited in a very character- 
istic way on boiling a weak solution of the phosphinous acid with acetate of zinc. 
Accordingly, the remainder of the barium salt was exactly decomposed with sulphuric 
acid, the solution filtered from the precipitated sulphate of barium and diluted. Acetate 
of zinc was then added (to the boiling solution), when it became turbid, and deposited a 
small quantity of a resinous salt. This was filtered off and excess of acetate of zinc, 
added, when a salt was precipitated (when the solution had been boiled for some time) in 
the characteristic form of the phosphinite, viz., minute droplets, with crystalline points 
projecting in all directions. A determination of zinc proved it to be the pure 
substance. 

Analysis. 

0-3044 gave 00670 ZnO=00537 Zn=l7-64 per cent. 

Obtained. Calculated for { (C 7 H 7 )P0 2 H } 2 Zn 

Zinc 17-64 17-33 



The results of these experiments show that in the reaction occurring between benzyl 
alcohol and the phosphorus mixture all the possible products of the oxidation of the 
three benzyl phosphines are produced, while in addition, in the experiment last described, 
iodide of tetrabenzyl phosphonium was also formed. In this experiment the conditions 
were slightly different from those of the preceding, for an excess of the phosphorus mixture 
was employed, but whether this influenced the course of the reaction or not we have not 
had the opportunity to decide. 

When we come to inquire into the mechanism (so to speak) of the reaction by which 
these different substances are formed, our attention is first arrested by the fact that 
two series of phosphorised derivatives are produced by the action of benzyl alcohol 
on the mixture of phosphorus and its iodide. One of these (excluding toluol) is 
represented by a solitary substance, viz., iodide of tetrabenzyl phosphonium, which 
contains less oxygen than the alcohol. The other includes all the possible products of 
the oxidation of benzyl phosphines, and, with the exception of oxide of tribenzyl phos- 
phine, contains more oxygen than benzyl alcohol. To a certain extent, therefore, the 
reaction is analogous to that which occurs between water and the phosphorus mixture, in 
which phosphonium iodide is a product of reduction, while phosphorous and phos- 
phoric acids are products of oxidation. To account for this double set of 
actions in the case of benzyl alcohol {i.e., reduction and oxidation) is not 
difficult. The primary reducing agents are phosphorus and its iodide, while in the later 
stages of the reaction there is hydriodic acid also. As to the oxidising agent, iodine is 
visibly liberated in the first phases of the reaction, and probably water also, so that we 
think it probable that it is by their interaction that some of the oxidised products are 



600 



PROF. LETTS AND MR R. F. BLAKE ON 



formed. It is, however, very difficult, if not indeed impossible, to gain a complete insight 
into the changes which occur. But, on the other hand, it is easy to write a set of 
equations to account for the formation of the different products. Thus, the following may 

be written: — 

(1) PI 2 +C 7 H 7 OH + H 2 = (C 7 H 7 )H 2 P0 2 +HI + I. 

(2) PI 2 + C 7 H 7 OH + 2H 2 + 1 = (C 7 H 7 )H 2 P0 3 + 3HI . 

(3) PI 2 +2C 7 H 7 OH =(C 7 H 7 ),HPO„+HI + I. 

(4) PI 2 +3C 7 H 7 OH + HI = (C 7 H 7 ) 3 PO + 2H 2 + 3I. 

(5) PI 2 +4C 7 H 7 OH + 4HI =(C 7 H 7 ) 4 PI + 4H 2 + 5I. 

(6) PI 2 +2H 2 

(7) PI 2 +4H 2 0+3I 

(8) 2HI + C 7 H 7 OH 

(9) P+I 2 



= H 3 P0 2 +HI + I. 
= H 3 P0 4 +5HI. 
= C 7 H 8 + H 2 + I 2 . 
= PL. 



Let us next compare the quantities of the different substances actually obtained with 
those required for the above reactions, remembering that the "balance sheet" will be 
extremely rough. In Experiment 3 the following were the quantities employed and 
produced : — 



Employed in the Reaction. 


Produced by the Reaction. 


Phosphorus, 


555 


Crude monobenzyl phosphinic acid, 


600 






„ dibenzyl „ „ 


53 5 






„ tribenzyl phosphine oxide, 


300 






„ toluol, 


450 


Iodine, ... . 


94-5 


Hydriodic acid (calculated), 
Phosphoric acid (calculated from 50 


952 






grms. BaHP0 4 ), 


21-0 


Benzyl alcohol, ..... 


1590 


Amorphous phosphorus, 


35 




309 




339-7 



The quantities of phosphorus, water, and benzyl alcohol, taking part in the reaction, may 
be accounted for as follows : — 



Phosphorus — 



Taken, 
Remaining, 



55 - 5 grms. 
350 „ 



Used up in the reactions, . . 20 - 5 „ 
The phosphorus used up in the reactions is accounted for thus — 



60 grms. (C 7 H 7 )H 2 P0 3 


contains 


10*8 phosphorus 


53-5 „ (C 7 H 7 ) 2 HP0 2 


?» 


6-7 


300 „ (C 7 H 7 ) 3 PO 


» 


2-9 


500 „ BaHP0 4 


!> 


6-6 



270 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 



607 



Water- 



Set free in reaction, 4 



3'4 ? grms, 







122 „ 


Required for reaction, 


1 . 

2 . 


? (traces) 
12-5 


5» 3J 


6 . 

7 . 


? (traces) 
15-4 

27-9 


Benzyl alcohol — 

Taken, . 




. 1590 


Accounted lor thus — 






Required for reaction, 1 . 
„ 2 . 


? (traces) 
37-6 


>) )> 


3 . 


470 


)) >3 


4 . 


30-3 


3J JJ 


5 . 

8 . 


? (traces) 
52-8 

167-7 



In glancing over the foregoing table, the following points become apparent : — 

(1) The quantity of products is of greater weight than the sum of the quantities of 
benzyl alcohol, phosphorus, and iodine employed. 

(2) The quantity of phosphorus used up in the reactions is less than that required 
for the weights of the different substances obtained. This is probably due to the following 
causes : — 

(A) The "unacted upon" phosphorus is not pure phosphorus, but contains organic 
compounds; therefore the real quantity used up was probably rather more than 20*5 
grms. (B) All the products were weighed in a slightly moist condition, hence their real 
weights are no doubt slightly less than those actually obtained, and this is also shown 
by the fact that the benzyl alcohol required for the reactions is about 5 per cent, higher 
than the amount used. (C) The whole of the phosphate of barium was not weighed, but 
its amount calculated, on the assumption that f of the total quantity were actually 
weighed. 

(3) The quantity of water set free (or supposed to be set free) is less than one-half of 
that which is required for those reactions in which it is supposed to take part. This is 
important, as, we think, it clearly points to the fact that at least some of the above 
equations do not represent the actual changes occurring, but that in all probability iodized 
derivatives are produced in the first instance, ivhich are subsequently decomposed by the 
water which is added. Thus, the difference in the water supposed to be set free and 
required in the equations (14*1 grms.), almost agrees with the amount required for the 

VOL. XXXV. PART II. (NO. 15). 5 E 



608 PROF. LETTS AND MR R. F. BLAKE ON 

production of the phosphoric acid (15*4 grms.). And that this supposition is a true one 
is also clearly shown by the rise in temperature which takes place when the product of 
the reaction is treated by water, and by the complete change which occurs in its 
appearance, as well as by the fact that the weight of products is greater than the collec- 
tive weights of phosphorus, benzyl, alcohol, and iodine originally taken. 

It is also quite possible that the whole of the tribenzyl phosphine oxide is formed by 
the action of caustic baryta on iodide of tetrabenzyl phosphonium produced in the first 
instance : but we have not decided this point. 

(4) The quantity of benzyl alcohol employed is less than the quantity calculated as 
being equivalent to the weights of the different substances obtained. The difference, 
however, is not great (about 5 per cent.), and is accounted for, no doubt, as we have 
explained. 

The reaction with benzyl alcohol is a very interesting one, and we believe that it is 
capable of extension to other alcohols. Some preliminary experiments which we have 
made justify us in this belief, and we intend to pursue the inquiry. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 60& 



PART III.— THE PRODUCTS OF THE OXIDATION OF BENZYL 

PHOSPHINES. 

The substances which we describe in this part of our paper are four in number, viz., 
benzyl phosphinous acid and beDzyl phosphinic acid, derived from the primary phosphine ; 
dibenzyl phosphinic acid, from the secondary phosphine ; and oxide of tribenzyl phosphine, 
from the tertiary base. # Benzyl phosphinous acid we obtained for the first time by the 
oxidation of the primary phosphine, while we discovered benzyl phosphinic acid during our 
investigations on the action of benzyl alcohol on a mixture of phosphorus and its iodide. 
One of us and W. Wheeler were the first to obtain dibenzyl phosphinic acid, which we 
isolated from the products of Hofmann's sealed tube reaction. In this part of our com- 
munication we describe the properties and compounds of the three acids, as well as those 
of the tertiary phosphine oxide, which we have so frequently, and at times unexpectedly, 
encountered in the course of our investigations. 

(1) Benzyl Phosphinous Acid, (C 7 H 7 )H 2 P0 2 . 

We have obtained this acid in three different ways — (1) By the oxidation of mono- 
benzyl phosphine ; (2) Among the products of Hofmann's sealed tube reaction ; (3) 
Among the products of the action of benzyl alcohol on a mixture of phosphorus and its 
iodide. We have already described the steps which we employed for isolating it in each 
of these three cases, but it may not be unadvisable to describe the best method of 
obtaining it. 

The primary phosphine is allowed to oxidise in the air, care being taken to prevent 
the temperature from rising too high. The final product is a thick syrupy liquid, which 
contains phosphoric acid and benzyl phosphinic acid, in addition to the phosphinous 
acid. Either of the two following methods may be employed for isolating the latter : — 

(1) The viscous mass is dissolved in water and the solution just neutralised with 
baryta. It is then filtered from the precipitated phosphate of barium and evaporated to 
a small volume, when practically the whole of the phosphinate of barium separates out. 
The phosphinous acid is then obtained by decomposing the filtered solution with the 
proper quantity of sulphuric acid. 

(2) The aqueous solution of the product of oxidation is dissolved in water and 
precipitated with acetate of lead, filtered from the phosphate and phosphinate of lead, 
which separates, precipitated with sulphuretted hydrogen, and the filtered solution 
evaporated until the whole of the acetic acid has passed off. 

Properties. — Benzyl phosphinous acid is a syrupy liquid which refuses to crystallise. 
It is fairly soluble in water, but separates from a strong solution in oily drops. It dis- 

* As we have mentioned on p. 585, it is possible that other oxidised derivatives exist, but the above are the most 
important. 



010 PROF. LETTS AND MR R. F. BLAKE ON 

solves easily in alcohol, and it is also soluble in ether. It is a monobasic acid, and most 
of its salts are very soluble. They are consequently troublesome to obtain pure, but 
the following were prepared and analysed. 

Barium Benzyl Phosphinite. — Obtained by neutralising the crude acid formed by 
the oxidation of the primary phosphine, filtering off the barium phosphate, boiling to 
precipitate the barium phosphinate, again filtering and concentrating to a syrup. The 
latter placed in vacuo solidified to a granular crystalline mass. This was squeezed as 
dry as possible in a cloth filter, and for analysis a sample was pounded up and exposed 
to the air until it ceased to lose weight. 



"■e* 



Analysis. 

1-1383 lost at 110° C. 01443 H 2 = 12-67 per cent. 
1-1383 gave 0-5119 BaSO 4 = 0-301 Ba = 26-44 „ 

Obtained. Calculated for (C 7 H 7 HP0 2 ) 2 Ba,4H 2 

Water, . . . 1267 1387 

Barium, . . . 26-44 2639 

The salt in all probability suffers a slight decomposition at 110° C, and also becomes 
partially oxidised. For after it had been heated and was dissolved in water, a distinct 
odour of the primary phosphine was observed. No doubt these facts account for the 
discrepancy between the observed loss at 110° C, and the calculated amount of water of 
crystallisation. 

Calcium benzyl phosphinite was obtained by neutralising a solution of the acid 
with chalk and evaporating the filtered solution. The salt separated out in crystalline 
crusts during the evaporation in a very characteristic manner. It appears to be less 
soluble in hot than in cold water. 

Analysis. 

•6943 lost at 110° C. 01 13 = 1-62 per cent. 
•6943 gave 0-1070 CaO = 0-76428 Ca = lL00 „ 

Obtained. Calculated for 3{(C 7 H 7 HP0 2 ) 2 Ca},H 2 

Water, .... 1-62 1-68 

Calcium, ... 1100 11-23 

Magnesium Benzyl Phosphinite. — This salt was prepared by boiling a solution of 
the acid with excess of carbonate of magnesium, filtering and concentrating the solution 
first in a water-bath, and afterwards in vacuo over sulphuric acid. When the solution 
grew syrupy, crystalline crusts were slowly deposited. 

Analysis. 

07354 lost (after 8 days' heating) at 110° C. 01538 H 2 = 20-91 per cent. 
0-7354 gave 01938 Mg 2 P 2 7 = 0041 9 Mg = 5'69 





Obtained. 


Calculated for (C 7 H 7 HP0 2 ) 2 Mg,5H 2 


Magnesium, 


569 


5-66 


Water, 


2091 


2122 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 611 

Zinc Benzyl Phosphinite. — Obtained by adding acetate of zinc in excess to a rather 
dilute solution of the acid and boiling for some time, when the compound was deposited 
as a granular crystalline powder. The salt is anhydrous. 

Analysis. 

(i.) 03044 gave 0-067 ZnO = 00537 Zn = 17 -64 per cent. 

(ii.) 05145 „ 01133 „ =009091 „ =17-66 

(m.) 0-8757 „ 01946 „ =0-15616 „ =1783 

Obtained. 

< * < Calculated for 

I. II. III. {(C 7 H 7 )HP0 2 } 2 Zn 

Zinc, . . 17-64 1766 1783 17-33 

Cadmium Benzyl Phosphinite. — This salt was prepared by decomposing a solution 
of the barium salt with the equivalent quantity of cadmium sulphate, filtering from the 
sulphate of barium and evaporating the filtrate. When the solution grew concentrated 
a white flocculent salt separated, which was filtered off and analysed. The mother-liquors, 
when further concentrated, gave a slimy deposit, and on cooling a crystalline salt. 

Analysis. 

0-4953 gave 01569 CdO = 0137287 Cd = 2771 per cent. 

Obtained. Calculated for {(C 7 H 7 )HP0 2 } 2 Cd 

Cadmium, . . . 27-71 26-54 

Lead Benzyl Phosphinite. — A rather concentrated solution of the pure phosphinous 
acid (obtained by decomposing the barium salt with sulphuric acid) was mixed with the 
calculated quantity of acetate of lead. The solution was filtered from a slight white 
precipitate which had been thrown down, slightly acidulated with acetic acid and warmed, 
when a colourless crystalline salt was precipitated. It was rapidly washed, dried, and 
analysed. 

Analysis. 

0-4269 (air-dried salt) lost at 110° C. 0-0061 H 2 0= 1-42 per cent. 
04269 gave 0-2446 PbS0 4 = 01671 Pb =3914 „ 

Obtained. Calculated for 2(C r H 7 HP0 2 ) 2 Pb,H 2 

Water, .... 1-42 114 

Lead, 39-14 39-35 

The salt fused when dried at 110° C, and smelt slightly of the primary phosphine. It 
dissolves readily in water. When its solution is concentrated by boiling it separates out 
in oily drops. 

Attempts to prepare Benzyl Phosphinite of Copper. — The behaviour of benzyl 
phosphinite of barium (or any other phosphinite) with acetate of copper is very character- 
istic. On mixing the two solutions a pale green precipitate is produced in a strong 
solution, but no precipitate in a weak solution. A specimen of the green precipitate was 
prepared and analysed. 



012 PROF. LETTS AND MR R. F. BLAKE ON 

Analysis. (The salt lost 7*43 per cent, on drying it at 110° C.) 
0-10273 gave 0"0304 CuO = 00242 Cu = 25"26 per cent. 

Calculated for 



Obtained. C 7 H 7 P0 3 Cu (C 7 H 7 HP0 2 ) 2 Cu 

Copper, . . . 2526 27-03 1700 

The salt was probably impure benzyl phosphinate of copper, and not phosphinite at all. 
The mother-liquors which had been filtered from it rapidly decomposed and a yellow to 
red precipitate separated, having the appearance of cuprous oxide. On warming the 
wash waters, especially if they contain excess of acetate of copper, the light green salt 
is precipitated, but redissolves as the solution cools. It appears, therefore, that benzyl 
phosphinite of copper, if it exists at all, is a very unstable substance, and rapidly oxidises 
to phosphinate. The behaviour of a phosphinite with a copper salt affords an excellent 
test for the acid. 

Action of Heat on Benzyl Phosphinous Acid. — A quantity of the acid was heated 
in a small retort. It rapidly decomposed, and a liquid distilled, having the characteristic 
odour of the primary phosphine, smoking on coming in contact with the air, and giving 
a crystalline compound with hydriodic acid, which volatilised in that gas in the charac- 
teristic manner of the phosphine hydrioclate. The reaction probably proceeds according 

to the equation — 

3(C 7 H 7 )H 2 P0 2 = 2(C 7 H 7 )H 2 P0 3 + (C 7 H 7 )H 2 P , 

giving the primary phosphine and benzyl phosphinic acid, and is analogous to the decom- 
position which phenyl phosphinous acid suffers when heated.* 

(2) Benzyl Phosphinic Acid, (C 7 H 7 )H 2 P0 3 . 

This acid is produced along with the benzyl phosphinous acid in the three reactions we 
have mentioned above, but it is only formed in very small quantity during the oxidation of 
the primary phosphine, benzyl phosphinous acid being, as we have mentioned, by far the 
chief product. A similar remark applies to the reaction occurring between oxide of zinc, 
benzyl chloride, and phosphonium iodide, only very small quantities of the benzyl phos- 
phinic acid being found among the products. The only satisfactory method for obtain- 
ing it is by the third process we have mentioned, i.e., the action of benzyl alcohol on a 
mixture of phosphorus and phosphorus iodide, when it is produced in large quantities, and 
is easily separated from the other products of the reaction. 

The following are the details of the process : — 5 5 '5 grms. of phosphorus are dissolved 
in an equal weight of bisulphide of carbon in a flask about 1 litre in capacity, and 9 4 '5 
grms. of iodine are gradually added. The bisulphide is then distilled off and the last 
traces removed by a current of dry carbonic anhydride. An upright condenser is now 
fitted to the flask (which must be filled with dry carbonic anhydride) and 159 grms. 
of benzyl alcohol cautiously added through the condenser from a tap funnel. As soon as 

* Michaelis and Ananoff, Ber., vii. 1688 ; and Kohler and Michaelis, Ber., x. 807. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 613 

the mixture of phosphorus and its iodide is moistened with the benzyl alcohol a violent 
reaction occurs — violet vapours of iodine appearing and a dense white smoke. It is 
advisable, as soon as the reaction has commenced, to run in the benzyl alcohol in fairly 
large portions, waiting, however, each time until it moderates. By this means constant 
ebullition is kept up until the end of the reaction. Towards the close of the operation 
(that is to say when all or most of the alcohol has been added) the flask is well shaken 
from time to time, so as to mix its contents thoroughly. This generally starts the reaction 
afresh, and ebullition becomes energetic. 

When all action is over, the product is allowed to cool, water is added until the flask 
is half full, and the mixture thoroughly shaken, when it becomes warm. Then a current 
of steam is blown through it and the vapours condensed, as they contain toluol, &c, the 
steaming being continued so long as volatile matters pass over. The aqueous solution is 
next decanted through a linen filter from the brown resinous mass which remains, and the 
latter is again mixed with water and steamed as before, the operation being repeated a 
third time. All the aqueous extracts are united and allowed to cool, when a considerable 
quantity of dibenzyl phosphinic acid separates out. This is filtered off through a linen 
filter and well squeezed, so as to avoid loss of liquid. The solution is next evaporated 
until fumes of hydriodic acid appear, and then allowed to stand for some hours, when the 
crude benzyl phosphinic acid separates out. It is thoroughly squeezed in a linen filter, 
and the mother-liquors again concentrated, &c, when a fresh quantity of the acid usually 
separates. 

The yield of crude acid from the quantities mentioned should be about 60 grms. The 
acid thus obtained is very impure, and contains considerable quantities of hydriodic and 
phosphoric acids. It may be recrystallised from water, but cannot be readily purified by 
that means, as the impurities cling most tenaciously to it. By far the best method of 
purification consists in converting it into its barium salt, and decomposing the latter with 
sulphuric acid. The barium salt is easily obtained pure by neutralising a dilute solution 
of the acid with caustic baryta, filtering from the precipitate of phosphate of barium and 
boiling the solution, when it separates in the crystalline state. 

Properties. — Monobenzyl phosphinic acid is a colourless body which crystallises from 
a hot and concentrated aqueous solution. It is readily soluble in water and alcohol. 
Its melting point was found to be 169°- 169° "5 (corr.). It is a dibasic acid, and forms a 
number of salts, most of which crystallise easily from water. 

For analysis a quantity of the acid was prepared from the barium salt, and was dried 
first in a desiccator then for a short time at 110° C. 

Analysis. 

( 01842 H 2 = -020466 H= 5-63 per cent. 
Odbdl gave j . 6485 C02 = . l76863 c =48 . 70 n 

* 0-2559 gave 0-1608 Mg 2 P 2 7 = -044908 P. = 17-54 

* By the oxide of copper method. 



614 PROF. LETTS AND MR R. F. BLAKE ON 





Obtained, 


!arbon, . 


4870 


[ydrogen, 


5-63 


hosphorus, . 


17-54 



Calculated for (C 7 H 7 )H 2 P0 3 
- 48-83 
5-23 
18-02 



Salts. 



Barium Benzyl Phosphinate (normal) (C 7 H 7 )BaP0 3 ,2H 2 0. — This is an extremely 
characteristic compound, and is readily obtained by neutralising a fairly strong solution 
of the acid with a warm solution of caustic baryta, when it crystallises out in thin plates. 
As it is much less soluble in hot than in cold water, it is precipitated on warming a weak 
solution, and if the heat is applied gradually (for instance by immersing the vessel con- 
taining the solution in a water-bath full of cold water and then heating the latter) it 
separates in large scales with the lustre of mother-of-pearl. This property furnishes an 
easy method for identifying the acid and separating it from other substances. 

Analysis. 

1-4246 (dried at 110° C.) gave 1-0809 BaS0 4 = 0*63555 Ba = 4461 per cent. 

Obtained. Calculated for C 7 H 7 P0 3 Ba 

Barium, .... 4461 44-62 

1-3643 lost at 110° C. 0-1389 = 10-18 per cent. 

Obtained. . Calculated for C 7 H 7 BaP0 3 ,2H 2 

Water, .... 10-18 1049 

Solubility. — An excess of the salt was allowed to stand for some days with water, 
well shaken repeatedly at intervals and the mixture filtered. Temperature of solution 
before filtering, 9°'7 C. 100 c.c. of this solution weighed (at 9°*7 C.) 101*446 grms., and 
left on evaporation 1*6115 grms. of the anhydrous salt =1*8005 grms. hydrated salt. 
Therefore, at 9°*7 C, 

100 c.c. of saturated solution contain 1-8005 grms. hydrated salt. 
100 grms. of water dissolve 1*807 „ „ 

Another quantity of the cold saturated solution was boiled for some time, then 
filtered through a hot- water funnel and cooled to 9 C *7 C. 

100 c.c. of this solution weighed (at 9°*7 C.) 100*475 grms., and left on evaporation 
0*3855 grin, of salt (dried at 110° C). Therefore, at 100° C. (or rather at the boiling 
point of the solution) 

100 c.c. of solution contain (about) 0'4307 grm. hydrated salt. 
100 grms. of water dissolve 0*4305 „ „ 

Acid Barium Benzyl Phosphinate, (C 7 H 7 . HP03) 2 Ba,3H 2 0. — This salt was obtained 
by adding to a solution of the normal salt the calculated quantity of benzyl phos- 
phinic acid and evaporating the mixture to a very small bulk, when, on cooling, the 
compound separated as a crystalline mass. It is very soluble in water. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 615 

Analysis. 

2-7733 grms. lost at 110° C. 0-2884 grm. = 10-39 per cent. 

2-7733 „ gave 1-2159 BaS0 4 = 0-7149 Ba = 25-77 

Obtained. Calculated for (C 7 H 7 HP0 3 ) 2 Ba,3H 2 

Water, .... 10-39 1013 

Barium, .... 25-77 2570 

Calcium Benzyl Phosphinate {normal), (C 7 H 7 )CaP0 3 ,H 2 0. — This salt was obtained 

by mixing a solution of the acid with a neutral solution of acetate of calcium and 

warming the mixture. The calcium salt was then precipitated in glittering scales, very 

similar in appearance to the barium salt. Prepared by this method it is thrown down 

on warming, even in weak solutions. The molecule of water which it contains is water of 

halhydration, as it is not lost at 110° C, and not completely at 200° C. 

Analysis. 

0-6177 lost at 200° C. 00428 H 2 = 6-92 per cent. 

I. 06459 (dried at 110° C.) gave 01563 CaO = 011164 Ca = l7'28 per cent. 
II. 0-6177 „ „ „ 0-1542 CaO = 0-11014Ca= 17-83 „ 

Obtained. 
, * v Calculated for 



I. II. (C 7 H 7 )CaPO.,,H 2 

Water, . . ... 6-92 7-89 

Calcium, . . 17-28 17-83 17*54 

Magnesium Benzyl-Phosphinate (normal), (C 7 H 7 )MgP0 3 ,H2(). — Was prepared in 
exactly the same way as the calcium salt, that is to say, by adding a solution of acetate 
of magnesium to the acid and warming the mixture. It is then precipitated, even in a 
fairly dilute solution, as a granular white powder. The water which it contains must be 
considered as water of halhydration, as it is only driven off at 200° C. 

Analysis. 

0-7858 (dried at 110° C.) lost at 200° C. 0-0683 H 2 = 8-69 per cent. 

1-0681 „ „ gave 0-5617 Mg 2 P 2 7 = 012144 Mg = 11-36 

Obtained. Calculated for (C 7 H 7 )MgP0 3 ,H a O 

Water, .... 8-69 8-49 

Magnesium, .... 1L36 11-32 

Zinc Benzyl Phosphinate, (C 7 H 7 )ZnP0 3 ,H 2 0. — Was prepared by adding acetate of 

zinc to a boiling solution of the acid, when it was thrown down as a white amorphous 

and bulky precipitate. On concentrating the mother-liquors a granular white powder 

separated, which was not examined. The molecule of water which the salt contains is 

driven off at 110° C. 

Analysis. 

0-7902 lost at 110° C. 0-0551 = 6-97 per cent. 

0-7902 gave 0-2544 ZuO = 0-20414 Zn = 25-83 „ 

Obtained. Calculated for (C 7 H 7 )ZnP0 3 ,H 2 
Water, .... 6"97 7*11 

Zinc, .... 25-83 25"69 

VOL. XXXV. PART II. (NO. 15). 5 F 





Obtained. 


Water, . 


5-60 


Cadmium, 


3705 



GIG PROF. LETTS AND MR R. F. BLAKE ON 

Cadmium Benzyl Phosphinate, (C 7 H 7 )CdP0 3 ,H 2 0. — A solution of the pure acid was 
largely diluted and mixed with a freshly prepared solution of acetate of cadmium. No 
precipitate was produced until the cadmium salt was added in considerable excess ; then 
however a salt was precipitated in minute spherical crystalline masses. On concentrating 
the mother-liquors only a trifling quantity of salt was obtained. 

Analysis. 

1-0181 lost (after 3 days' heating) at 200° C. 00571 H 2 = 5 60 per cent. 
10181 gave 04312 CdO = 03773 Cd = 37-05 

Calculated for (C 7 H 7 )P0 3 Cd,H 2 
600 
37-33 

Lead Benzyl Phosphinate, (C 7 H 7 )PbP0 3 ,H 2 0. — Was obtained as an amorphous white 

precipitate on boiling a solution of the acid with acetate of lead. The molecule of water 

which the salt contains is only lost at a temperature of 200° C. 

Analysis. 

1-4939 lost at 200° C. 00341 H 2 =228 per cent. 
1-4939 gave 11748 PbS0 4 = 0-8025 Pb = 537l „ 

Obtained. Calculated for (C 7 H 7 PbP0 3 ) 2 ,H 2 

Water, .... 2-28 2"33 

Lead, .... 5371 53'62 

Copper Benzyl Phosphinate, (C 7 H 7 )CuP0 3 ,H 2 0. — On mixing a solution of the acid with 
acetate of copper a light blue precipitate of the copper salt is thrown down. 
Analysis. 

0939 lost (after 4 days' heating) at 200° C. 00571 = 6-08 per cent. 

(On the 4th day the salt began to decompose.) 
0939 gave 0-2973 CuO = 0-23708 Cu = 25'24 per cent. 

Obtained. Calculated for (C 7 H 7 )CuP0 3 ,H 2 

Water, .... 608 7 17 

Copper, .... 25-24 25"09 

Silver Benzyl Phosphinate. — This salt is thrown down as a bulky white precipitate on 
mixing a neutral solution of the potassium salt with nitrate of silver. It darkens when 
warmed, and is unstable. It was not analysed owing to this fact. 

Potassium Benzyl Phosphinate was prepared by neutralising the acid with caustic 

potash solution and evaporating to small bulk, when eventually a crystalline mass 

remained. The salt is very soluble both in water and alcohol. 

Analysis. 

J lost at 110° C, 00278 H 2 = 6-86 per cent. 

I gave 0-7392 K 2 PtCl 6 =011815 K=2919 „ 





Obtained. 


Calculated for (C 7 H 7 )K 2 P0 3 ,H 2 


Potassium, 


. 2919 


29-32 


Water, 


6-86 


676 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 617 

Sodium Benzyl Phosphinate and Ammonium Benzyl Phosphinate resemble the 
potassium compound, and were prepared in a similar manner. They were not analysed. 

Action of a Moderate Heat on Benzyl Phosphinic Acid. — When moderately heated 
the acid loses water, the amount finally lost varying apparently with the conditions of 
the experiment. In one experiment about five grms. of the acid* were placed in a 
distilling flask, through which a slow current of hydrogen passed, and the flask was 
heated in an oil-bath for some hours at a temperature ranging from 180°-195° C, then 
to 220°-230° C. for nearly a week. During this time a trace of the acid or some product 
of its decomposition appeared to volatilise. At the end of this time the loss amounted 
to 4*19 per cent. The equation — 

2C 7 H 7 H 2 P0 3 = H 2 + (C 7 H 7 ) 2 H 2 P 2 5 

requiring 5 '23 per cent. The product was light brown in colour, and very crystalline. 
On boiling with water it appeared to be quite insoluble. A little of it, mixed with a 
solution of caustic baryta, gave at once a crystalline deposit, which seemed to be very 
insoluble. The whole of the product was pounded in a mortar and boiled with water, 
in which some, but not all dissolved. The solution, which was fairly dilute, was then 
mixed with a slight excess of caustic baryta and gradually heated. Long slender 
needles separated, which were well washed and air dried. 

Analysis. 

1-2391 lost at 240° C. 01173 = 9-46 per cent. 

1-2391 gave 0-5733 BaS0 4 = 0-337 Ba = 27-19 
06822 „ 08146 CO 2 = 0'222163 0=32-56 
06822 „ 0-2363 H 2 = 0-026255 H= 384 

Water, ..... 
Barium, ..... 
Carbon, ..... 
Hydrogen, ..... 

In another experiment about 5 grms. of the acidt were heated at once to 230° C, 
and kept at that temperature for a considerable time. The final loss amounted to 14 '4 4 
per cent. The equation — 

2(C 7 H 7 )H 2 P0 3 ,H 2 = 3H 2 + (C 7 H 7 ) 2 H 2 P 2 5 

requires a loss of 14*21 per cent. 

It would appear from this result that the two specimens of acid experimented with 
were different, the latter containing water of crystallisation, the former none. The 
product from this second experiment was broken up, pounded fine, and boiled with water, 
in which it ultimately dissolved. The solution was divided into three parts, A, B, and C. 

* Not absolutely pure, but recrystallised two or three times from the crude product. 

t A very pure specimen of the acid, which had been obtained from the barium salt, and was snow-white. 



Obtained. 


Calculated for (C 7 H 7 ) 2 P 2 5 Ba,3H 2 


9-46 


10-48 


2719 


26-60 


32-56 


32-62 


3-84 


3-88 



618 PROF. LETTS AND MR R. F. BLAKE ON 

A was slightly evaporated, and on cooling gave a crop of colourless crystals, unlike 
the original acid, and probably the pyro acid. 

B was evaporated to a small bulk, and was completely converted into the original 
acid, as was proved by the production of the characteristic barium salt. 

C was neutralised with caustic baryta, and gave the long needles of the pyro- 
phosphate exactly similar to those obtained in the first experiment. 

We may conclude, then, that at a temperature of 200°-230° C, benzyl phosphinic acid 
is converted into the pyro acid, (C 7 H 7 ) 2 H 2 P 2 5 (which gives a characteristic barium salt, 
and is less soluble than the original acid). The pyro acid, on boiling for a long time with 
water, is rehydrated, and gives the original benzyl phosphinic acid.* 

Action of a Rapid Heat on Benzyl Phosphinic Acid. — 5 grms. of the acid were cau- 
tiously heated in a small retort over the naked flame of a Bunsen burner. At first a few 
drops of water passed over, then a yellowish oil distilled, and presently a sudden 
decomposition occurred, the mass charring. The lamp was instantly removed, but the 
decomposition went on spontaneously. 

The receiver contained 1 to 2 c.c. of a yellowish oil, smelling very slightly of the 
primary phosphine, and partly solidifying on cooling. The black tarry matter in the 
retort contained a little free phosphorus. It was boiled with water and the solution 
filtered. The filtrate was neutralised with caustic baryta, and gave a white amorphous 
precipitate, which, when filtered off and dissolved in nitric acid, gave a strong phosphoric 
acid reaction with molybdate of ammonium. 

The filtrate from the amorphous precipitate, when boiled, gave characteristic crystals 
of barium benzyl phosphinate, showing that some of the acid had escaped decomposition. 
According to Michaelis and Matthias (loc. cit.), phenyl phosphinic acid decomposes at 
250° C. into benzol and meta-phosphoric acid, with charring — 

C c H 5 H 2 P0 3 = C 6 H 6 +HP0 3 . 

They do not mention how the benzol was identified. 

We had not sufficient of the oily distillate to determine its boiling point or to 
identify it as toluol, but as the production of phosphoric acid was proved, it is probable 
that benzyl-phosphinic acid decomposes like the phenylated acid thus — 

C 7 H 7 H 2 P0 3 =C 7 H 8 +HP0 3 . 

Action of Phosjihorous Acid on Benzyl Phosphinic Acid. — Having obtained the 
benzyl phosphinic acid in large quantities, and with ease, we were anxious to discover some 
method by which it could be reduced to the phosphine, and as we had proved that the 
action of heat alone did not lead to the desired result, we decided to try the effect of 

* Michaelis and Matthias (Ikr., vii. 1070) and Michaelis (Annalen, 181, p. 323) found that the corresponding 
phenyl phosphinic acid, when heated to 200° C, loses sufficient water to give the pyro acid analogous to the one we 
oMained, viz., (C G H 5 ) 2 H 2 P 2 6 ; while at 210° C. three molecules of the acid lose two molecules of water, giving, they 
suppose, (C 6 H s ) 3 H 2 P 3 O r . They did not succeed in isolating either of the pyro acids or in obtaining their salts, as they 
re-hydrate when treated with water. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 619 

heating the acid with crystalline phosphorous acid, in the hope that a reaction would 
occur according to the following equation : — 

SHgPOs+CyH^POs^SHgPO.+CyH^P, 

which would be analogous to the reaction which phosphorous acid suffers when heated 

alone — ■ 

4H 3 P0 3 = 3H 3 P0 4 + PH 3 . 

Accordingly, 5 grms. of benzyl phosphinic acid were mixed with a large excess 
(20 grms., the calculated quantity being only 7"1 grms.) of crystallised phosphorous 
acid, and the mixture heated in a distilling flask. It soon fused to a clear liquid, and 
then a little water came off. Presently the mass began to disengage phosphuretted 
hydrogen, and to froth considerably. After some time an oily liquid distilled, which was 
colourless, and finally decomposition and charring occurred. The whole of the liquid 
distillate was redistilled, and gave 1'3 c.c. of an oil having the smell and properties of the 
primary phosphine. The reaction seems then to proceed in the desired manner, and if 
the whole of the liquid obtained consisted of the phosphine, the yield was nearly 50 per 
cent, of the theoretical quantity. In all probability other phosphinic acids would be 
reduced by phosphorous acid, and, if so, the action is of some importance, as we believe 
that no other method for their reduction has yet been discovered. 

Action of Pentachloride of Phosphorus on Benzyl Phosphinic Acid. — We were 
anxious to obtain the two chlorides, (C 7 H 7 )C1HP0 2 and (C 7 H 7 )C1 2 P0. 

10 grms. of the benzyl phosphinic acid and 14 grms. of pentachloride of phosphorus 
were mixed in a distilling flask, when a pretty brisk action occurred spontaneously, the 
mixture growing very hot and torrents of hydrochloric acid coming off. As soon as the 
action had moderated, the mixture was heated, and 8 grms. of liquid were obtained dis- 
tilling below 110° C. The residue in the flask was allowed to cool, and hardened into a 
brown viscous mass. It was then heated, when the thermometer rose slowly to 300° C, 
during which a colourless liquid distilled. There remained in the flask a black liquid, 
which solidified to a hard resin, and which contained free phosphorus. The distillate 
boiling from 110°-300° C. was in too small a quantity to redistil, but it probably con- 
tains one of the two chlorides, for when left for some time in contact with the air, it was 
converted into a crystalline mass, presumably benzyl phosphinic acid. The reaction was 
not sufficiently definite to invite further investigation. 

(3) Dibenzyl Phosphinic Acid, (C 7 H 7 ) 2 HP0 2 . 

We have obtained this acid by the second and third reaction described at the beginning 
of this part of our paper, and also by fusing the tertiary phosphine oxide with caustic 
potash. To isolate it from the products of the reaction occurring between phosphonium 
iodide, zinc oxide, and benzyl chloride, the contents of the sealed tubes, after having been 
steamed with water to drive off the primary phosphine, are boiled for a long time with 



620 FROF. LETTS AND MR R. F. BLAKE ON 

caustic potash or caustic baryta solution. This is then decanted off and acidulated with 
hydrochloric acid, when dibenzyl phosphinic acid is precipitated, but in a very impure 
condition, and only a small quantity is obtained. The reaction occurring between 
benzyl alcohol and a mixture of phosphorus and its iodide yields it in large quantities, 
and with ease. We have on p. 612 described this reaction, and have explained how the 
crude dibenzyl phosphinic acid is separated. 

In order to purify it, the crude product is thoroughly washed with water and 
recrystallised from boiling alcohol. 

Properties. — Dibenzyl-phosphinic acid is a colourless substance, almost insoluble in 
water, but fairly soluble in hot alcohol. It crystallises from the latter solvent in very 
thin iridescent plates, which, when dry, have a mother-of-pearl lustre. Its melting point, 
determined with an Anschutz thermometer, was found to be 192° C. Heated above its 
melting point, it decomposes, but also partly volatilises unchanged. Heated with 
pentachloride of phosphorus, it appears to decompose, and to give among other products 
chloride of benzyl. It forms a number of very characteristic salts, many of which are 
beautifully crystalline. 

Analysis* 

r 1-4058 C0 2 = 0-3834 C = 67-99 per cent. 
•5639 gave \ 03315 H 2 6 = 0-036833 H= 6-53 
1 0-2639 Mg 2 P 2 7 = 0073701 P = 13-06 

Obtained. Calculated for (C 7 H 7 ) 2 HP0 2 

Carbon, 67'99 68-29 

Hydrogen, 6'53 6-09 

Phosphorus, .... 13-06 12-60 

Salts. — 

Barium Dibenzyl Phos%)hinate, {(C 7 H 7 ) 2 P0 2 } 2 Ba,8H 2 0. — This salt was obtained by 
neutralising a hot and concentrated solution of caustic baryta with the acid, filtering 
from a few insoluble flakes,t and allowing the solution to cool, when very thin plates 
separated with the lustre of mother-of-pearl. The mother-liquors, on spontaneous 
evaporation, gave large and very beautiful thin plates, radiating from a common centre. 

Analysis. 

(1) 0-4637 grms. lost at 110° C, 0-0855 = 18-43 per cent. 

(2) 1-1392 „ „ „ 0-2112 =18-53 
11392 grms. gave 0-3439 BaS0 4 = 02022 Ba. = 1774 

Obtained. 

Calculated for 
{(C 7 H 7 ) 2 P0 2 } 2 Ba,8H 2 



I. 


ii. 


1853 


18-43 


17-74 





Water, . . . 18-53 18-43 ' 18'67 

Barium, . . . 1774 ... 17-76 

The salt effloresces in dry air. It is soluble in alcohol. 

* The combustion was made with pure oxide of copper, and the phosphorus determined as described at p. 556. 
t We do not know to what these are due. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 621 

Calcium Dibenzyl Phosphinate, {(C 7 H 7 ) 2 P0 2 [ 2 Ca,8H 2 0. — This compound was pre- 
pared by warming the acid with a little water and slaked lime, until the solution was 
neutral. The filtered solution gave on cooling thin plates very much like the barium 
salt. On boiling the crystals with their mother-liquors they grew opaque and insoluble, 
no doubt from loss of water. 

Analysis. 

1-0637 lost at 110° C, 0-2227 = 20"93 per cent. 

1-0637 gave 0-0902 CaO = 0-0644 Ca= 605 „ 

Obtained. Calculated for {(C 7 H 7 ) 2 P0 2 } 2 Ca,8H 2 

Water, . . 2093 21-36 

Calcium, . . 6-05 5-93 

Magnesium Dibenzyl Phosphinate, {(C 7 H 7 ) 2 P0 2 } 2 Mg,3H 2 0. — This salt is very 
characteristic. It was prepared by slowly warming a dilute solution of the potassium 
salt with acetate of magnesium, when it was slowly deposited in colourless needles of 
considerable length. 

Analysis. 

06422 lost at 110° C, 00563 = 8-76 per cent. 

0-6422 gave 0-1301 Mg 2 P 2 7 = 0-0281 Mg = 437 

Obtained. Calculated for {(C 7 H 7 ) 2 P0 2 } 2 Mg,3H 2 

Water, . . 8-76 9-50 

Magnesium, . 4-37 4-22 

Cadmium Dibenzyl Phosphinate, {(C 7 H 7 ) 2 P0 2 } 2 Cd. — Was obtained by adding sul- 
phate of cadmium to a solution of the potassium salt, when an amorphous white preci- 
pitate was thrown down, which rapidly became crystalline. 

Analysis. 

08052 gave 0-1757 CdO = 0-1537 Cd = 1908 per cent. 

Obtained. Calculated for (C 7 H 7 ) 2 P0 2 } 2 Cd 

Cadmium, . . . 19-08 I860 

The mother-liquors, on evaporation, gave a considerable quantity of a crystalline 
salt. 

Copper Dibenzyl Phosphinate, {(C 7 H 7 ) 2 P0 2 } 2 Cu. — This body was prepared by adding 
sulphate of copper to a solution of the potassium salt, when a blue amorphous precipitate 
came down. This was washed, air-dried, and analysed. 

Analysis. 

10037 gave 0-1482 CuO = 01183 Cu = 11-78 per cent. 

Obtained. Calculated for {(C 7 H 7 ) 2 P0 2 } 2 Cu 

Copper, . . . 11-78 11-47 

The mother- liquors when boiled gave a green precipitate, and the blue salt, on simply 



(y-22 PROF. LETTS AND MR R. F. BLAKE ON 

standing with water, also seemed to become green. The green compound was not inves- 
tigated. 

Silver Dibenzyl Phosphinate, (C 7 H 7 ) 2 P0 2 Ag. — Was prepared by dissolving some of 
the acid in alcohol and adding strong aqueous nitrate of silver solution, when the silver 
salt was precipitated in very fine colourless needles. These blackened slightly when 
dried. 

Analysis. 

0-3122 grm. gave 0125 AgCl = 0-094087 Ag = 3013 per cent. 

Obtained. Calculated for (C 7 H 7 ) 2 P0 2 Ag 

Silver, . . . 30-13 3059 

Sodium Dibenzyl Phosphinate, 2{(C 7 H 7 ) 2 P0. 2 Na},7H 2 0. — Prepared by neutralising 
caustic soda solution with the solid acid and evaporating to a small volume, when the salt 
crystallised out in plates, and was easily recrystallised from boiling water. It is readily 
soluble. 

Analysis. 

(1) 13114 lost at 110° C, 02492 = 19-00 per cent. 

(2) 1-3999 „ „ 0-2660 = 19-00 „ 

Obtained. 
' * » Calculated for 



I. II. 2{(C 7 H 7 ) 2 P0 2 Na},7H 2 

Water, . . 19-00 19-00 19-03 

Potassium Dibenzyl Phosphinate, 2{(C 7 H 7 ) 2 P0 2 K},7H 2 0. — Prepared in the same 
manner as the sodium salt, and has similar properties. 

Analysis. 

1-2054 lost at 110° G, 0-2213 =18-35 per cent. 

1-2054 gave 0*860 K 2 PtCl 6 =0*1374 K = 11-39 

Obtained. Calculated for { (C 7 H 7 ) 2 P0 2 K } 2 ,7H 2 

Water, .... 18-35 18-15 

Potassium, . . . 11-39 1123 

Ammonium Dibenzyl Phosphinate, (C 7 H 7 ) 2 P0 2 NH 4 ,7H 2 0. — This salt was prepared 
by neutralising the acid with ammonia. The solution, when highly concentrated, solidified 
to a crystalline mass. 

Analysis.* i 

0-9105 gave 0-2339 Pt=00427 NH 4 = 468 per cent. 

Obtained. Calculated for (C 7 H 7 ) 2 P0 2 NH 4 ,7H 2 

Ammonium, . . . 4 - 68 4*62 

* The salt was distilled with caustic potash and the ammonia which was evolved absorbed by hydrochloric acid 
and precipitated as chloroplatinate, the latter being subsequently ignited, and the ammonia calculated from the weight 
of platinum left. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 623 

Action of Heat on Dibenzyl Phosphinic Acid. — Preliminary experiments with the 
pure acid in a test-tube seemed to show that it distilled unchanged. A quantity was 
heated in a distilling flask. It boiled above the boiling point of mercury, and the 
distillate on cooling formed a semi-solid crystalline mass, smelling very slightly of the 
primary phosphine. As the distillation proceeded, the residue in the flask darkened. 
The whole of the distillate was warmed with caustic potash, in which most of it seemed 
to dissolve, but a crystalline residue remained saturated with some oily matter. The 
solution was filtered from this residue, precipitated with excess of hydrochloric acid, and 
the washed precipitate recrystallised from alcohol. It separated in characteristic thin 
plates of dibenzyl phosphinic acid, and was identified by its melting point and other 
properties as that body. The crystalline matter insoluble in potash was well washed 
with water and then crystallised from hot alcohol, when it separated in needles, having 
the melting point of tribenzyl phosphine oxide, and it was further identified as that body 
by the production of the characteristic bromine compound. 

The residue in the distilling flask was black and tarry, and was dissolved when boiled 
with caustic soda. The solution thus obtained, when acidulated with hydrochloric acid, 
gave a crystalline precipitate (probably crude dibenzyl phosphinic acid), and when filtered 
from this, was found to contain abundance of phosphoric acid. 

We may conclude from the above results that when dibenzyl phosphinic acid is heated 
it volatilises in a great measure unchanged, but that a portion decomposes, giving 
phosphoric acid, tribenzyl phosphine oxide, and toluol — 

2(C 7 H 7 ) 2 HP0 2 = (C 7 H 7 ) 3 PO + HP0 3 + C 7 H 8 . 

Action of Pentachloride of Phosphorus on Dibenzyl Phosphinic Acid. — 10 grms. of 
the acid and 10 grms. of pentachloride of phosphorus* were placed in a distilling flask, and 
the latter gradually treated in a paraffin-bath. A reaction soon occurred, and torrents of 
hydrochloric acid gas came off. As soon as the action had moderated the product was 
distilled. 7 grms. of liquid came off before 110° C, then the thermometer rose slowly to 
300° C, without remaining constant at any temperature, during which about 7 grms. of 
an oily liquid came off. By this time the residue in the distilling flask was very syrupy 
and of a brown colour, and on admitting air it smoked, and seemed to be giving off free 
phosphorus. 

The two distillates, when redistilled, passed over in a great measure below 110° C. 
They were shaken with water, and after decomposition an oily liquid floated on the surface 
(in all about 2 grms.), which was unmistakably benzyl chloride. We could not isolate 
any acid chloride, (C 7 H 7 ) 2 P0C1, and a profound decomposition seems to occur, possibly 
according to the equation — 

(C 7 H 7 ) 2 HP0 2 + 3PC1 5 = 2PCLO + HC1 + 2PC1 3 + 2C 7 H 7 C1 . 

* The quantities required for the equation (C 7 H 7 ) 2 HP0 2 +PCl 5 = (C 7 H 7 ) 2 POCl + HCl+POCl 3 are 10 grms. of the 
acid and 8'5 grms. of pentachloride of phosphorus. 

VOL. XXXV. PAST II, (NO. 15). 5 G 



&24 PROF. LETTS AND MR R. F. BLAKE ON 



(4) Tribenzyl Phosphine Oxide, (C 7 H 7 ) 3 PO . 

This body has been obtained by a number of different methods, among which are the 
following : — 

(1) Action of benzyl chloride on phosphonium iodide (Fleissner).* 

(2) From the products of the action of benzyl chloride on phosphide of sodium 

(Letts and Collie) .1' 

(3) From the products of Hofmann's sealed tube reaction (between benzyl chloride, 

zinc oxide, and phosphonium iodide) (Letts and Blake)4 

(4) From the products of the action of benzyl alcohol on a mixture of phosphorus 

and phosphorus iodide (Letts and Blake). § 

(5) From the products of the action of benzyl alcohol on phosphonium iodide 

(Ledermann).|| 

(6) By the action of alkalies on salts of tetrabenzyl phosphonium (Letts and 

Collie).1T 

(7) By the action of heat on hydrate of tetrabenzyl phosphonium (Letts and 

Collie). ## 

(8) By the oxidation of the tertiary phosphine (Letts and BLAKE).tt 

(9) By the action of fused potash on dibenzyl phosphinic acid (Letts and Blake). tt 

Properties. — Crystallises in colourless highly refractive needles. These are some- 
times thick, at others quite fine and silky. The body is soluble in alcohol, bisulphide of 
carbon, glacial acetic acid, and chloroform ; almost insoluble in water and ether. When 
heated it melts at 216°-216°'5 C. (corr.), and sublimes at a higher temperature with 
considerable decomposition. 

It combines with hydracids, halogens, and chloride of platinum to form unstable and 
possibly indefinite compounds, but its compound with iodide of zinc is stable and definite, 
and so is its nitro body, its sulphonic acid, and some other derivatives. 

Bromide, either 7(C 7 H 7 ) 3 PO,5Br 2 or 5(C 7 H 7 ) 3 PO,4Br 2 . — This compound is highly 
characteristic, and its production affords a ready means of identifying the oxide. It is 
readily obtained by adding bromine to a solution of the oxide in glacial acetic acid, and 
it crystallises in orange needles, which, when seen under the microscope, are found to 
consist of aggregations of minute rhombic plates. It may be recrystallised from a warm 
solution in glacial acetic acid (to which a little bromine is added). It is unstable, and 
loses bromine when boiled with water or glacial acetic acid. It is instantly decomposed 

* Fsleisner, Berichte, xiii. (1880) 1665. + Letts and Collie, these Transactions, xxx. part 1, p. 202. 

J Letts and Blake, this communication, p. 554. § Letts and Blake, this communication, p. 599. 

|| Lkdermann, Berichte, xxi. (1888) 405a. IT Letts and Collie, loc. cit., p. 196. 

** Letts and Collie, loc. cit., p. 215. t+ Letts and Blake, this communication, p. 574. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 



625 



by caustic potash, and the oxide re-formed. The bromine is, in fact, very loosely combined 
— as loosely as water of crystallisation is in ordinary salts. 

The compound has been repeatedly analysed, and we think it advisable to give the 
full results : — 



Analysis. 



Bromine, (1) 
(2) 
(3) 
(4) 
(5) 
(6) 
(7) 
(8) 
(9) 
(10) 

(11) 

(2) 
Carbon, 565 
Hydrogen, 5*3 



Bromine, 
Carbon, . 
Hydrogen, 



Obtained. 
26'0 Prepared with a hot solution of the oxide. 
^fc>-4< „ „ „ „ 

27-2 „ cold 

26-7 

26 - 6 „ „ „ „ 

26 - 64 Prepared with a cold solution of the oxide. 
26"85 Recrystallised from acetic acid. 
26-58 

] Prepared with a boiling solution of the oxide in 
| glacial acetic acid, and dried in vacuo. 



(4) 

56-4 

5-6 



(12) 

56-5 

50 



(9) 

56-9 

4-9 

Calculated for 



7(C 7 H 7 ) 3 PO,5Br 2 
26-31 
58-00 
483 



5(C 7 H 7 ) 3 PO,4Br 2 
28-5 
56-3 

4-7 



Chloride. — Obtained by passing chlorine gas into a solution of the oxide in warm 
acetic acid to saturation, and crystallised as the solution cooled in pale yellow crystals, 
having the appearance of pentachloride of phosphorus. The compound is very unstable, 
and loses chlorine in vacuo. 



Analysis. 

Chlorine, 

(It is probable that on drying the compound for analysis some chlorine was lost.) 



Obtained. 
120 



Calculated for 7(C 7 H 7 ) 3 PO,5Cl 2 
13-68 



Iodide. — Prepared by mixing hot solutions of iodine and the oxide in glacial acetic 
acid. Crystallises in minute red crystals of the colour of ferricyanide of potassium. 



Analysis. 



Iodine, 



Obtained. 
36-86 



Calculated for 7(C 7 H 7 ) 3 PO,5l 2 
3618 



* Analyses 1-5 were made by Letts and W. Wheeler, and from 6-8 by Letts and Blake, and the preparations 
obtained in both cases from Hofmann's "dibenzyl phosphine" {see p. 554). Analyses 9-11 were made by Letts and 
Collie, the preparation being made from the oxide obtained from the products of the action of sodium phosphide on 
benzyl chloride. 



frlti PROF. LETTS AND MR R. F. BLAKE ON 

Hydrochlorate* — When warmed with an aqueous solution of hydrochloric acid, the 
oxide remains unchanged, but if a current of the dry gas be passed over the finely 
powdered oxide, it is readily absorbed, and a compound is obtained which is unchanged 
in the air, but decomposes when heated or boiled with water into its constituents. 

Its formula appears to be 4(C 7 H 7 ) 3 P0,3HC1. 

Hyd rob r ornate. — According to Collie (loc. cit.), a similar compound is formed with 
hydrobromic acid. One of us and W. Wheeler obtained a compound by saturating a 
hot solution of the oxide in glacial acetic acid with hydrobromic acid gas, when it 
separated as the solution cooled in colourless crystals. Its composition varied widely in 
two separate preparations. 

Analysis. 

Obtained. Calculated for 



I. II. 4(C 7 H 7 ) 3 PO,3HBr 3(C 7 H 7 ) 3 PO,2HBr (C 7 H 7 ) 3 PO,HBr 

Bromine, 195 to 205 16-3 15-75 14*26 19*95 

Hydriodate. — Obtained by one of us and W. Wheeler by the same method as the 
hydrobromate. 

Analysis. 

Obtained. Calculated lor 



I. ii. 4(C 7 H 7 ) 3 PO,3HI 3(C 7 H 7 ) 3 PO,2HI (C 7 H 7 ) 3 PO,HJ 

Iodine, 2P0 21-5 22-9 20-88 28-34 

Platinum Salt. — Prepared by mixing alcoholic solutions of the oxide and chloride of 

platinum, and crystallised in minute orange leaflets. It is somewhat unstable, and its 

composition varies with the conditions under which it is obtained. Its probable 

formula is — 

4(C 7 H 7 ) 3 PO,2HCl ) PtCl 4 ,t but possibly 3(C 7 H 7 ) 3 PO,PtCl 4 + 

Palladium Salt. — Fleissner obtained this body by precipitating a solution of the 
oxide with palladous chloride. It is a brown-red crystalline mass of the formula 
3(C 7 H 7 ) 3 PO,PdCl 2 . 

Ferric Chloride Compound. — Sulphur yellow prisms of considerable size. Formula : 
3(C 7 H 7 ) :J PO,Fe 2 Cl 6 (Fleissner). 

Mercuric Chloride Compound. — Beautiful colourless prisms or pyramids. Formula : 
3(C 7 H 7 ) 3 PO,HgCl 2 (Fleissner). 

Cobaltous Chloride Compound. — Blue needles. Formula: 3(C 7 H 7 ) 3 PO,CoCl 2 
(Fleissner). 

* Collie, Ohem. Hoc. Jour., 1889, p. 227. t Letts and Collie, loc. cit. j Fleissner, loc. cit. 



BENZYL PHOSPHINES AND THEIR DERIVATIVES. 627 

Zinc Iodide Compound. — Obtained by mixing alcoholic solutions of the two bodies, 
and separated from the fairly concentrated solution in tufts of colourless plates of char- 
acteristic form. Formula : 2(C 7 H 7 ) 3 P0 5 ZnI 2 (Letts and Collie). 

Acetyl Chloride Compound (Collie). — Obtained by slowly evaporating a solution of 
the oxide in glacial acetic acid and chloride of acetyl, as a crystalline compound. 
Formula: (C 7 H 7 ) 3 PO.(CH 3 .COCl). It is unstable, and is decomposed into its consti- 
tuents by heat : when treated with boiling water and when mixed with an alkali. 

Sulphur Compound (Letts and Collie). — When the oxide is fused with sulphur a 
reaction occurs, which apparently varies with the temperature, and with the quantity of 
sulphur employed. If much sulphur is taken and the mixture heated to a high 
temperature, sulphuretted hydrogen is evolved, the mass becomes dark coloured and 
resinous products are formed. But if the proportion of sulphur is low (one molecule of 
the oxide to two atoms of sulphur), and the temperature kept at 240° C, the sulphur 
dissolves, no gas is evolved, and the product dissolves completely in a large quantity of 
boiling alcohol. The solution deposits on cooling long silky needles of a light buff colour 
of the composition, 5(C 7 H 7 ) 3 PO,S. The melting point was found to be 211°-212° C. 
(uncorrected). 

Nitro Compound. — Obtained by Collie by dissolving the oxide in cold sulphuric 
acid, adding excess of nitric acid, and pouring the mixture into water. Prepared by us 
by dissolving the oxide in cold fuming nitric acid and pouring the solution into water.* 
White amorphous mass soluble in glacial acetic acid, melting at 100° C, and deflagrating 
at a higher temperature. Formula : (C 7 H 6 (N0 2 )} 3 PO. It is unchanged when boiled 
with chromic acid, but is oxidised to para-nitro benzoic acid (and presumably phosphoric 
acid) when warmed with an akaline solution of permanganate of potash (Collie). 
Attempts by Dr Collie and ourselves to obtain the corresponding amido body were 
unsuccessful. 

Sulphonic Acid (Collie). — The oxide, when dissolved in strong sulphuric acid, does 
not react unless the temperature is raised above 100° C. ; between 150° C. and 170° C, 
the whole of the oxide is readily converted into a sulphonic acid, (C 7 H 6 S0 3 H) 3 PO. This 
acid is soluble in water, and can be obtained pure from its barium salt. It is semi- 
crystalline, and dries to a syrup over sulphuric acid. Monoacid Barium Salt, 
{ (C 7 H 6 S0 3 ) 2 Ba (C 7 H 6 )S0 3 H}PO. — Obtained by neutralising the acid with caustic baryta. 
Is uncrystallisable. Silver Salt. — White flocculent precipitate. Copper Salt. — Green, 
and soluble in water. Lead Salt. — Soluble. 

The acid behaves with oxidising agents like the nitro body. 

Action of Fused Caustic Potash on the Oxide. — When the oxide is heated with caustic 

* The discovery of the nitro compound was made independently by Dr Collie and ourselves. 
VOL. XXXV. PART II. (NO. 15). 5 H 



628 PROF. LETTS AND MR R. F. BLAKE ON BENZYL PHOSPHINES. 

potash or soda, it fuses and floats on the surface of the melted alkali. No violent action 
occurs, but on cooling the mixture and treating it with water, the greater portion dissolves, 
and acids then precipitate dibenzyl phosphinic acid. The folio wiug reaction therefore 
occurs : — 

(C 7 H 7 ) 3 PO + KHO = (C 7 H 7 ) 2 P0 2 K + C 7 H 8 . 

Singularly enough, when dibenzyl-phosphinic acid is heated by itself, the oxide is 

formed — 

2(C 7 H 7 ) 2 P0 2 H = (C 7 H 7 ) 3 PO + HP0 3 + C 7 H 8 . 



lH.e Roy. Soc. Edm T 



Vol. XXX\ 



PROFESSOR LETTS and M ? R.F.BLAKE ON BENZYL PHOSPHINES 





,rn fi h ". 



( 629 ) 



XVI. — On the Anatomy, Histology, and Affinities of Phreoryctes. By Frank E. 
Beddard, M.A., Prosector to the Zoological Society of London, Lecturer on 
Biology at the Medical School of Guy's Hospital. (With a Plate.) 

(Read 18th February 1889.) 

Introductory. 

Some time since I received from Mr W. Smith of Ashburton, New Zealand, a single 
example of a worm belonging to the little known genus Phreoryctes, and published in 
the Annals and Magazine of Natural History for June 1888, a short account of the 
reproductive organs of this worm, for which I proposed the name of Phreoryctes 
Smiihii. 

More recently (November 1888) I have received, through the kindness of the same 
gentleman, a large number of specimens of the same species ; the study of these enables 
me to offer to this Society a more complete account of the structure of Phreoryctes, 
which is at present very imperfectly known. It will be seen, however, that the present 
paper, though more extended, and dealing with organs to which little or no reference 
was made in my former paper, contains no important correction of facts stated in that 
paper. I have also appended (p. 638) a discussion of the systematic position and 
affinities of Phreoryctes. 

Historical. 

The genus Phreoryctes was first described by W. Hoffmeister, in a paper upon the 
terrestrial Annelida of Germany [A 4, # p. 193] ; the specimen was discovered by Menke, 
and Hoffmeister gave it the name of Haplotaxis Menkeana. In this paper the 
worm is compared to Gordius, on account of the thickness of the integument. 
Hoffmeister regards it as intermediate between Gordius and Lumbricus. It is 
stated (erroneously) that the outer rows of setse have vanished ; the inner rows of setse 
consist of simple setse, which are longer than the setse of Lumbricus, and, instead of being 
hooked at the free extremity, are only slightly bent. A figure is given (pi. ix. fig. 
vii.) of the head, showing the division of the prostomium. 

In his work, "Die bis jetzt bekannten Arten aus der Familie der Regenwiirmer" 
[A 5], Hoffmeister alters the name Haplotaxis to Phreoryctes, since the former name 
had been already applied to a plant. In that work the genus Phreoryctes is thus 
denned : — " Zwei Reihen einzelner, gerader Borsten ; riisselformige Oberlippe ; Giirtel 
fehlt ; kein Vulva." 

The species Phreoryctes Menkeanus is briefly re-described and figured. 

The next contribution to our knowledge of this genus is by Schlotthauber [A 9]. 

* These numbers refer to the List of Literature on p. 639. 
VOL. XXXV. PART II. (NO. 16). 5 I 



630 ME FRANK E. BEDDAED ON THE 

His paper on the subject is reviewed by Leuckart [A 6j. Schlotthauber proposed to 
alter the name to Georyctes, since the worm lives chiefly in damp soil, and but rarely in 
water. The species described is regarded as different from Phreoryctes Menkeanus, and 
is named Georyctes Lichtensteinii. 

Claparede, in one of his memoirs upon Oligochseta [A 2, p. 27 5 J, gave a brief descrip- 
tion, with illustrations, of a worm, which he termed Nemodrilus filiformis, being 
evidently unaware of its identity with Phreoryctes. It is perfectly clear, however, from 
his figure of the head of this worm [A 2, pi. iii. fig. 16], that it agrees with Phreoryctes 
in the division of the prostomium, and indeed the fact is mentioned in the text. Atten- 
tion is drawn to the fact that the setae are isolated, and arranged in four rows ; the dorsal 
setae are stated to be twice as long as the ventral setae, and are figured. The nephri- 
diopores are placed in front of the ventral setae. 

The earliest paper which gives anything like a complete account of the structure of 
Phreoryctes Menkeanus is by Leydig [7]. This paper also contains the first description 
of the genital organs, which are stated to be three pairs of glands situated in segments 
IX.— XL A fourth pair, however, is figured, as was pointed out by Timm. 

An important contribution to the anatomy of Phreoryctes is that of Timm [A 10]. 
This naturalist was the first to investigate the worm by the section method, and he was 
able therefore to add something to Leydig's facts. He confirms the statement of the 
latter that the gonads are in the IXth, Xth, Xlth, and Xllth setigerous segments {i.e., 
in the Xth, Xlth, XTIth, and XHIth, as usually reckoned), and describes three pairs of 
spermathecae in segments VII., VIII., and IX. Timm remarks upon the "Hirudinean" 
character of the longitudinal muscle fibres, which had been only up to that time de- 
scribed in the Enchytraeida (by Ratzel) among the Oligochseta. The genus Phreoryctes 
is treated of by Vejdvosky [A 11] in his System und Morphologie der Oligochceten, pp. 
48-50, where an account is given of the previous literature of the subject. Vejdovsky 
describes the species Phreoryctes filiformis which had been already noticed by himself, 
and which he considers to be the same species as those described by Schlotthauber, 
Claparede, and Noll. 

Dr F. Noll [A 8] described some specimens of this genus which were collected in the 
year 1835 at Riidesheim, on the Rhine, by C. von Heyden. The specimens are to be 
found in the Museum of the Senckenberg Society of Naturalists, labelled Lumbrico- 
gordius Hartmanni. The name Lumbricogordius is evidently given on account of the 
view as to the affinities of the genus held by Hoffmeister. Noll names the species 
Phreoryctes Heydeni ; it was met with by him at St Goar, on the banks of the Rhine. 
His paper is illustrated by a plate. 

The next contribution to the anatomy of Phreoryctes is by myself [A 1]. 

The paper contains a brief description of the reproductive organs of a new species of 
the genus which is named after its discoverer, Phreoryctes Smithii. 

Finally, there are two notes by Giard [A 3] upon the occurrrence of Phreoryctes 
Menkeanus in the north of France, at Douai, and at Boussac (Creux). Giard remarks 



ANATOMY, HISTOLOGY, AND AFFINITIES OF PHREORYCTES. 631 

in the former paper that the apparent rarity of Phreoryctes Menkeanus is possibly due 
to its having been confounded with Gordius. 

A paper by G. da Rossi [A 12] I have been unable to see. 



Description of Phreoryctes Smithii, F. E. B. 
Mode of Life. 

The first specimen of Phreoryctes Smithii which Mr Smith forwarded to me was found 
in marshy soil coiled into a ball with a number of others. The second lot of specimens 
were found by that gentleman in a forest pool ; in the bottle which contained these were 
a quantity of smaller worms, which had much the same appearance as the Phreoryctes. 
Mr Smith conjectured that they might be young individuals of that species. They turn 
out to be a species of the genus Limnodrilus, but I have not yet determined whether they 
constitute a new species of that genus, or whether they are referable to one of the forms 
already known. 

Phreoryctes Smithii, therefore, like the other species of the genus, can live in water 
or in moist earth. 

The specimens from the forest pool were very much larger than those which I received 
at first ; the length of the longest specimen was about eight inches and the thickness 
one-fifteenth of an inch. They were nearly all sexually mature, and had a well-developed 
clitellum. 

External Characters. 

The large size of this species, both as regards length and breadth, appear to distinguish 
it from both Phreoryctes Menkeanus and Phreoryctes filiformis. It is true that the 
former species has been found to grow longer, but its thickness is inconsiderable. The 
measurements of Phreoryctes filiformis are given by Vejdovsky [A 11, p. 49], as 10 to 
14 cm. in length by f to 1 mm. in breadth. My former statement therefore [A 1, p. 
389], that this species is not so elongated as either of the other two species at present 
known, is justified by the study of these larger individuals. 

Setce. 

In the other two species of Phreoryctes — viz., Phreoryctes Menkeanus and 
Phreoryctes filiformis — the setae have been described as showing a very remarkable 
arrangement ; of both of these species it has been stated that each segment possesses 
only four setae, disposed in four rows ; when, as very occasionally happens, one of these 
setae is supplemented by another, the latter is to be looked upon asa" reserve seta " 
according to Vejdovsky [A 11]. 

Phreoryctes Smithii can at once be distinguished from these two species by the 
arrangement and number of the setse. ^They commence, as in all Oligochaeta, on the 



632 MK FRANK E. BEDDARD ON THE 

second segment, and are disposed in four longitudinal rows ; but there are invariably two 
closely approximated setae in both the dorsal and ventral rows. The second seta does not 
in any way suggest the idea of its being a " reserve seta ;" indeed, in several cases 
" reserve setae," more or less immature, and to the number of two, were present in addition 
to the fully formed and functional setae. 

The setae are implanted so as to divide the circumference of the body into four areas ; 
the distance, however, between the ventral pairs is less than that between the dorsal and 
ventral pair of one side, and the distance between the two dorsal pairs. In fact, the 
diagrammatic transverse section illustrating my previous note upon this worm [A 1, pi. 
xxiii. fig. 7], proves to be not quite correct when checked by my subsequent observations 
upon other examples. The arrangement of the setae, therefore, in Phreoryctes Smithii 
brings the genus nearer to the Lumbriculidae, when there are four rows of pairs of setae. 

Another point in which Phreoryctes Smithii differs from the other species of the 
genus is in the shape of the setae. 

In both the other species the shaft, the portion implanted in the body-wall, is 
straight; this is shown in the figures, e.g., of Timm and Claparede and Vejdovsky. 
In Phreoryctes Smithii this is not the case ; the setae are bent throughout, and have the 
shape so characteristic of the setae of the Oligochaeta (vide Plate, fig. 6). 

In their shape, therefore, as well as in their number and arrangement, the setae of 
Phreoryctes Smithii are much more like those of the Lumbriculidae than are the setae of 
Phreoryctes Menkeanus and Phreoryctes filiformis. On the other hand, the present 
species agrees with the others in the fact that the dorsal setae are much larger than the 
ventral. Fig. 6 is drawn with the aid of the camera lucida, and therefore illustrates 
accurately the relative sizes of the dorsal and ventral setae. This difference is not, 
however, obvious on the anterior segments of the body. 

Absence of Genital and Penial Seta. — It is important to put on record the fact, that 
in the neighbourhood of the reproductive apertures there is no modification whatever 
of the setae. It is not possible to state with absolute certainty that Phreoryctes has no 
genital setae, but it is at least highly probable that this is the case. This again is a 
point which bears upon the affinities of the genus. Most Oligochaeta show some modifica- 
tion of the setae on the genital segments, but this is apparently not so with, e.g., the 
Lumbriculidae, which family Phreoryctes, as has been already pointed out, resembles in 
other particulars. 

Body -Wall. 

Phreoryctes Smithii does not differ from the other species of the genus in the 
structure of the body-wall ; I have, therefore, but little to say under this head. 

The epidermis (see fig. 12) is like that of Lumbricus and many other Oligochaeta 
in the specialisation of its cells into large glandular cells and narrow interstitial packing 
cells. 

The relative thickness of the two muscular layers is shown in fig. 12. 



ANATOMY, HISTOLOGY, AND AFFINITIES OF PHREORYCTES. 633 

Clitellum. 

In the majority of my specimens the clitellum was developed. The position of the 
clitellum in this genus has not been hitherto known ; and, as this organ is of some little 
importance in the classification of Oligochseta, it is particularly desirable to have some 
information upon the point. The clitellar region was obvious, in all the specimens which 
had reached that degree of maturity, by its swollen distended appearance and whitish 
colour. The swollen appearance and the white colour are, however, due not so much to 
the modification of the integument in this region of the body, as to the mass of generative 
products, principally spermatozoa, which are developed in these segments, and cause them 
to be considerably distended. The comparatively slight increase of thickness in the 
epidermis of the clitellum, as compared with the epidermis over the general body-surface, 
is not sufficient to distinguish this part of the body when examined without the aid of a 
microscope. When the body is slit open, and the integument examined microscopically, 
the extent of the clitellum is quite obvious ; it extends over three complete segments 
and the part of a fourth ; the posterior boundary of the clitellum coincides with the 
furrow separating segments XIII.-XIV.; anteriorly the clitellum is not so sharply 
defined ; it commences on segment X. at or near the setse. 

The clitellum of Phreoryctes Smithii, therefore, occupies three segments and a portion 
of a fourth, commencing on the Xth, and ending at the posterior border of the XHIth 
segment. 

It forms a complete girdle round these segments, i.e., there is no ventral space not 
invaded by the glandular modification of the integument. 

The clitellum, therefore, includes all the apertures of the generative ducts, which are 
thus in the strictest sense " intraclitellian." 

The minute structure of the clitellum was investigated by longitudinal sections as 
well as by an inspection of the entire clitellar area mounted in Canada balsam unstained. 
Preparations of the latter kind showed that the epidermic cells were greatly modified 
upon the clitellum ; the cells were filled with rouDded, highly refractive bodies ; in 
sections the cells were seen to be considerably longer than the epidermic cells elsewhere ; 
but there was no trace in my preparations of more than a single layer of cells. It is 
important to notice this fact, because, if anything could be said in favour of the old 
division of Oligochseta into Oligochceta limicolce and Oligochce taterricolce, it is that the 
structure of the clitellum is markedly different in the genera assigned to the one group 
from what it is in genera assigned to the other group. In earthworms it appears to be the 
rule, and as far as we know at present, without any exceptions, that the clitellar epidermis 
consists of two more or less distinct strata ; on the contrary, in all the aquatic and other 
genera which have been referred to the Liniicolse, the clitellar epidermis only consists of 
one layer. I take this opportunity, however, of pointing out this structural difference 
between the " Limicolse " and " Terricolse," as I have not noticed any prominent statement 
of the facts as an argument in favour of the above mentioned classificatory scheme. 



634 MR FRANK E. BEDDARD ON THE 

This distinction may not, however, be so important as it at first sight appears to be ; 
and in any case we have as yet no information as to the condition of the clitellum in 
Moniligaster, a genus which shows so many important resemblances to certain Limicolous 
genera, though undoubtedly an earthworm. In the minute structure of its clitellum, 
Phreoryctes seems to agree pretty closely with Rhynchelmis, as described by 
Vejdovsky [B 3 J. 

Vascular System. 

The vascular system of Phreoryctes Smithii consists, as in the other two species of 
the genus, of a dorsal and ventral trunk, which are united by transverse branches. 

Fig. 7 illustrates the arrangement of the vascular trunks in some of the posterior 
trunks. It will be seen that the dorsal and ventral vessels are here connected in each 
segment by a transverse pair of trunks which do not pass straight from the dorsal to the 
ventral vessel, but have a very sinuous course. The coils of these lateral vessels are not, 
however, so complicated as in the anterior segments, and their calibre is also less. These 
vessels, moreover, are not invested by a thick sheath of peritoneal cells like the anterior 
lateral trunks. The vascular trunks of these posterior segments are precisely like those 
of Phreoryctes jiliformis, as described and figured by Vejdovsky [A 11, pi. xii. figs. 
8, 9]. In Phreoryctes Menheanus the lateral trunks are only connected with the ventral 
vessel ; they arise from the latter, and pass round the circumference of the ccelom for a 
considerable distance, but do not join the dorsal vessel. 

Reproductive Organs. 

In my previous paper upon the anatomy of Phreorytes Smithii, I was able to describe 
the testes and the ovaries as well as the efferent ducts. The specimen described in that 
paper was not fully mature, and I could find therefore no trace of the sperm-sacs and 
ovisacs, of which an account is given in the present paper. 

The testes are, as stated before, two pairs situated in segments X. and XL; they are 
not, however, of a simple conical form in the fully developed worm, but prolonged into 
several processes ; the digitate shape is due to the rapid and unequal proliferation of the 
testicular cells. The genera Lumbricus and Allolobophora (among others) have been 
distinguished by the form of the testes, which have been figured as of conical form in the 
one, and digitate in the other ; it is very possible that this difference does not really exist, 
but that it is merely due to the stage of development at which the organs have been 
studied ; such a difference occurs at any rate in Phreoryctes. 

Vasa Deferentia. 

There are two pairs of these ducts, which open independently of each other on to 
segments XL and XII. 



ANATOMY, HISTOLOGY, AND AFFINITIES OF PHREORYCTES. 635 

I find that I have not described their actual orifices with complete accuracy in my 
former note upon this species. 

The first pair of vasa deferentia do indeed open near to the ventral pair of setse 
between these and the dorsal pair, but the second pair are a little different. The vas 
deferens of each side is much shorter, and opens well in front of the ventral pair of setse 
of the Xllth segment, though behind the groove which separates this segment from the 
Xlth ; that there is really this somewhat unexpected difference between the two pairs 
of vasa deferentia I have been able to prove by longitudinal sections, which are much 
better than transverse sections for demonstrating such a point. In preparation of the 
worm, mounted entire in Canada balsam, some of the orifices of the sexual ducts were 
quite conspicuous. One of three such preparations which I have shows the external 
pore of the oviducts, and of the posterior pair of vasa deferentia ; in all three cases the 
orifices are situated on a line with and in front of the ventral setse ; the oviducal 
pores are placed further forward than the male pore — in fact, on the intersegmental 
furrow. 

Although out of the three specimens, in which I have examined the vasa deferentia, 
only one showed with perfect plainness the apertures of both pairs, the position of the 
anterior and posterior pairs respectively in each of the two remaining specimens leaves 
no doubt upon my mind that fig. 1 of the Plate is a correct representation of the position 
of the generative pores. 

The terminal section of the vas deferens is lined with a chitinous membrane ; this 
fact is not mentioned in my former paper ; I remark upon it here as it is a further 
point of similarity between the vasa deferentia and oviducts, which latter are also 
furnished with a chitinous lining distally. In my former paper I was only able to 
urge the great probability of the genus Phreoryctes being typically provided with four 
pairs of gonads and four pairs of independent ducts. I am now able to state this fact 
with certainty. 

The adult ova (fig. 11) are of very large size; the relation between their diameter 
and that of the body of the worm is shown in the figure cited. A' comparison of this 
section with a corresponding one of Clitellio arenarius [Beddard, B 1] shows that in 
that species the ova are of the same relative size as in Phreoryctes ; they are therefore 
actually smaller in Clitellio. The larger size of the ova of Phreoryctes is due to the 
great abundance of yolk, which is present in variously-sized spherules. The ovum is 
bounded externally by a fine homogeneous membrane ; it possesses a larger deeply 
staining nucleus, within which is a nucleolus. Vejdovsky [B 4] has recently published a 
careful description of the ripe ovum of Rhynchelmis. He remarks that the size of the 
mature ova vary according to the size of the individuals in which they are found, the 
largest worms producing the largest ova. I have pointed out myself that the relative 
size of the mature ova of Clitellio and Phreoryctes is about the same ; it is not, therefore, 
a distinction of any importance that the mature ova of Phreoryctes are larger than those 
of the Tubificidse (Clitellio). 



636 MR FRANK E. BEDDARD ON THE 



Affinities of Phreoryctes. 



Before discussing the affinities of Phreoryctes to other Oligochseta, it is necessary to 
inquire the relations of the present species to Phreoryctes Menkeanus and Phreoryctes 
filiformis ; this it is a little difficult to do satisfactorily, inasmuch as our knowledge of 
the reproductive organs of these two species is very meagre ; and generic and family 
distinctions are chiefly based upon variations of the reproductive system. With regard 
to Phreoryctes jiliformis, Vejdovsky [A 11] is unable to make any statements whatever 
about the reproductive organs ; he did not discover a trace of these organs in that 
species. Of Phreoryctes Menkeanus we know something about the gonads and the 
spermathecse ; the only investigators who have discovered these organs in Phreoryctes 
Menkeanus are Leydig [A 7] and Timm [A 10] ; according to Timm, there are three 
pairs of spermathecae situated respectively in segments VI., VII., and VIII.; these organs 
seem to be distinguished by their extraordinarily thick muscular walls. Both observers 
agree that there are four pairs of gonads, considered to be testes, situated in segments IX., 
X., XL, XII. Vejdovsky [A 11] has remarked that in all probability these supposed 
testes are in reality both testes and ovaries. Both Leydig and Timm failed to find any 
generative ducts ; and Timm supports the view of Leydig that the nephridia of the 
genital segments serve as efferent ducts. Although these descriptions are incomplete, 
they are, I think, sufficient to point to the conclusion that my species is a congener of 
Phreoryctes Menkeanus. It is so very rare among the Oligochseta for there to be four 
pairs of gonads (two pairs of testes and two pairs of ovaries), that this fact of agreement 
alone appears to demonstrate the generic identity of the two forms. 

The difference in position of the gonads, commencing as they are stated to do in the 
IXth segment in Phreoryctes Menkeanus and in the Xth segment in Phreoryctes 
Smithii, is due to a difference* in the way of enumerating the segments of the former 
species. 

The reason why neither Leydig nor Timm succeeded in discovering the reproductive 
ducts, may have been the extreme difficulty of seeing these structures in immature 
specimens. I have already pointed out that the funnels are made up of only one layer 
of cells closely pressed against the septum, and as the duct opens close to the posterior 
side of the same septum, there is no great length of vas deferens to attract the eye in 
transverse or longitudinal sections. It is perhaps more probable that in the specimens 
studied by Leydig and Timm they were even more rudimentary than in the specimen 
originally described by myself. That this was the case is rendered likely by the 
occurrence of nephridia in these segments, which, as we know from the discoveries of 
Vejdovsky, are present, in aquatic Oligochseta, in the genital segments before the 
generative ducts make their appearance ; when the vasa deferentia and oviducts are 
developed the nephridia of their segments atrophy and eventually disappear. 

* In this paper I count the peristomial segment as the first segment. 



ANATOMY, HISTOLOGY, AND AFFINITIES OF PHREORYCTES. 637 

The difference in the number of the spermathecse may well be only of specific value, 
although such a difference is not usual among the aquatic Oligochseta. 

These facts, coupled with the division of the prostomium by a transverse furrow, 
which is not found in any other Oligochseta that I am aware of, though it appears to be 
met with in the Oapitellidse, and the complicated structure of the longitudinal muscle 
bundles (cf. Plate, fig. 12, with Timm [pi. x. fig. 12]), lead me to regard the generic 
identity of the present species with Phreoryctes Menkeanus as established beyond any 
reasonable doubt. 

The salient facts in the anatomy of the genus Phreoryctes are then the following : — 

1. The body is extremely elongated, sometimes reaching a foot in length, while the 
diameter is very small. 

2. The prostomium is divided into two by a transverse furrow. 

3. The setae, are simple, not bifid; they are disposed in four rows of single setse, or of 
pairs of setse. 

# 4. There are no genital or penial setce. 

# 5. The clitellum occupies 3 to 4 segments from the Xth to the XHIth ; its epidermis is 
formed by a single layer of glandular cells, differing from the epidermis of the general 
body-surface by their glandular character and greater length. 
6. The structure of the longitudinal muscles (Plate, fig. 12). 

# 7. The nephridia commence in the sexually mature worm in the XVIth segment. 

8. There are two pairs of testes in the Xth and Xlth segments. 

9. Two pairs of vasa deferentia opening in to the exterior upon the Xlth and Xllth 
segments in front of the ventral setse, and opening by wide, simple (not plicated) funnels 
into the Xth and Xlth segments. 

10. Two pairs of ovaries in segments XII. and XIII. 

11. Two pairs of oviducts opening into the exterior on segments XIII. and XIV. 
near to the lower line between these segments, and the one in front on a line with the 
ventral setse. The oviducts open by wide simple funnels into the interior of segments 
XII. and XIII. The structure of these organs is exactly that of the vasa deferentia. 

# 12. Both series of ducts have the distal region lined with a chitinous membrane 
indicating (?) their origin from an ectodermic invagination. 

# 13. The developing spermatozoa are contained in sperm-sacs, which occupy 
segments 9 to 14 (about). 

# 14. The ova, which are when adult of very large size, undergo their development 
in egg-sacs, which occupy segments 14 to 16 (about). 

15. The spermathecse are present to the number of from two to three pairs in 
segments VII. , VIII. (and IX.). 

These characters distinguish the genus from any that is known, and quite justify its 
place as the type of a distinct family, to which position it is assigned by Claus and 
Vejdovsky [A 11]. Its relation to other families of Oligochseta are not, however, so clear. 

* The asterisks signify that the statements to which they are appended are made for the first time in the present paper. 
VOL. XXXV. PART II. (NO. 16). 5 K 



638 MR FRANK E. BEDDARD ON THE 

Vejdovsky [A 11, p. 1G3] considers that it is most nearly related to the Tubificidge 
(Limnodrilus). This conclusion was arrived at without any knowledge of the reproductive 
ducts which were at that time unknown ; their discovery renders that particular view of 
the affinities of Phreoryctes less satisfactory than it would otherwise be. 

It is also important to notice that in some respects the reproductive ducts of the 
adult Phreoryctes recall those of immature forms of other Oligochseta. For example, in 
Tubifex, the development of whose reproductive organs has been carefully worked out by 
Vejdovsky [A 11], the atrium, which forms so important a part of the adult efferent 
system, is very inconspicuous to begin with. In the young Tubifex [A 11, pi. ix. fig. 
18] it is represented by a very small dilatation, and the vas deferens itself is a short 
straight tube without any trace of the complicated windings which characterise the vas 
deferens of the sexually mature Tubifex. I have, myself, pointed out that in Clitellio 
arenarius [B 1] the vas deferens is at first short and straight, and afterwards becomes 
much longer and coiled several times, though its windings are never so complex as those 
of Tubifex. Psammoryctes [Vejdovsky, A 11, pi. x. fig. 17] is another instance to the 
point. 

A comparison between the vasa deferentia of the adult Phreoryctes and those of the 
young Tubifex, Clitellio, &c, is not vitiated by the suggestion that the resemblance is a 
purely superficial one, brought about by the fact that the more complicated vasa deferentia 
of Tubifex and Clitellio must necessarily pass through a stage such as that described ; 
that it is merely "the way they develop," and is void of any significance, just as some 
naturalists have attempted to show that the Nauplius is of no phylogenetic significance, 
inasmuch as a Crustacean with nineteen pairs of appendages developing progressively 
must pass through a stage in which there are only three pairs. 

In the case of the vasa deferentia there does not appear to be any a priori reason 
why the course of development should be as described, unless it has a phylogenetic 
significance. The nephridia of Tubifex, whose development Vejdovsky has also worked 
out, form a series of loops nearly as complicated as those of the adult before the con- 
nection with the distal vesicle is established [cf A 11, pi. ix. figs. 1, 2], 

Bergh's [B 2] observations upon the developing nephridia of Criodrilus are, as regards 
this point, confirmatory of the statements of Vejdovsky ; the nephridia arrive at some 
degree of complication before the external orifice is formed. Arguments of this kind, 
however, will not, in my opinion, apply to the atrium ; the appearance of this organ in 
many Oligochseta before the development of the vasa deferentia, and its almost universal 
presence and large size, indicate that it is a very characteristic organ in this group. Its 
full development, therefore, in Phreoryctes is not to be compared with the full develop- 
ment of the atrium in the immature Tubifex, but is due to degeneration. 

The reproductive organs of Phreoryctes are, therefore, to be regarded as being on the 
whole archaic in characters, but to have undergone some modification in the nearly 
complete disappearance of the atria. 

It follows, from what has been said in the foregoing pages about the reproductive 



ANATOMY, HISTOLOGY AND AFFINITIES OF PHREORYCTES. 639 

organs of the Oligochseta, that Phreoryctes stands midway between earthworms and the 
majority of those forms which were formerly grouped together as Limicolse. It has 
retained the paired testes and vasa deferentia of the former, while the vasa deferentia 
are shortened ; the paired ovaries and oviducts of Phreoryctes are represented in a few 
earthworms, but not in the lower aquatic Oligochseta. 

The clitellum again is more extensive than is the case with the lower forms, though 
as in them it consists of but a single layer of cells. 

The paired setse with simple hooked (not bifid) extremities are characteristic of the 
higher Oligochseta and of Phreoryctes. 

The structure of the body-wall in Phreoryctes, as regards the longitudinal muscular 
layer, is more like that of earthworms. 

On the other hand, the sperm-sacs and ovisacs as well as the large size of the ova, 
recall the Tubificidse and other families of aquatic or limicolous Oligochseta. 

Phreoryctes must therefore be regarded as occupying a position at about the middle 
of the series ; it has retained some of the characteristics of earthworms, but has in other 
respects acquired the simpler structure of the aquatic Oligochceta. 

It is, in fact, one of those forms which render a division of the Oligochseta into two 
groups, " Limicolce " and " Terricolcs," impossible ; it could not be definitely referred to 
either. 



LIST OF MEMOIRS REFERRED TO. 
A UPON PHREORYCTES. 

1. Beddard, F. E. On the Reproductive Organs of Phreoryctes, Annals and Magazine of Natural 

History, June 1888. 

2. Claparede, Ed. Recherches sur lAnatomie des Oligochetes, M&moires de la Soci4U de Physique 

et d'Histoire Naturelle de Geneve, t. xvi. (1862). 

3. Giard, A. Fragments Biologiques X. Sur une nouvelle station de Phreoryctes Mcnkeanus, 

Hoffmeister (Eaux de Source de Douai), Bulletin Scientifique de la France et de la Belgiquc, 
iii e ser., i re Annee (1888), p. 298, and idem, 2 e Ann^e (1889), p. 171. 

4. Hoffmeister, W. Beitrage zur Kenntniss deutschen Landanneliden, Archiv fur Natur- 

geschichte, Bd. i. (1843). 

5. Hoffmeister, W. Die bis jetzt bekannten Arten aus der Familie der Regenwurmer, 4to, 

Braunschweig, 1845. 

6. Leuckart, R. Berichte liber die Leistungen in der Naturgeschichte der niederen Thiere 

wahrend des Jahres 1859, Archiv fur Naturgeschichte, Bd. 26, 1860, p. 117. 

7. Leydig, F. Uber Phreoryctes Menkeanus nebst Bemerkungen, &c, Archiv fur miJcroscopische 

Anatomic, Bd. i. (1865), p. 249. 

8. Noll, F. C. Uber einen neuen Ringelwurm des Rheins, Archiv fur Naturgeschichte, Bd. xl. 

(1874), p. 260. 



640 ANATOMY, HISTOLOGY, AND AFFINITIES OF PHREORYCTES. 

9. Schlotthauber. Beitriige fiir Helminthologie, Amtliche Bericht uber 31 te Versammlung deutscher 
Naturforschcr und Artzte zu Gottingen, 1861. 

10. Timm, R. Beobachtungen an Phreoryctes Menkeanus, Hoffm. und Nais, Arbeiten aus dem zoolog- 

zootom. Institute in Wiirzburg, Bd. vi. (1883). 

1 1. Vejdovsky, F. System und Morphologie der Oligochaeten, Prag. 1884, p. 48. 

12. Da Rossi, J. Phreoryctes Menkeanus, Jahresbericht der Zool. Section des westfalischen pro- 

vinzial Vereins fiir Wissenschaft und Kunst, xv. p. 29. 



B UPON OTHER OLIGOCH^TA. 

1. Beddard, F. E. On certain Points in the Structure of Clitellio (Olaparede), Proceedings of the 

Zoological Society of London, 1888. p. 485. 

2. Bekgh, R S. Zur Bildungsgeschichte der Excretionsorgane bei Criodrilus. Arbeiten aus der 

Zool. Zoot. Institute in Wiirzburg, Bd. viii. p. 223. 

3. Vejdovsky, F. Anatomische Studien an Ehynchelmis limosella, Zeitschrift fiir wissenschaftliche 

Zoologie, Bd. xxvii. (1876). 

4. Vejdovsky, F. Entwickelungsgeschicbtliche Untersuchungen, Heft, i., Prag. 1888. 



EXPLANATION OF PLATE. 



Fig. 1. Lateral view of Phreoryctes Smithii (anterior segments), to illustrate the principal external 

characters, cf, orifices of spermatheca ; £, male ; £, female generative apertures ; 

cl, clitellar segment. 
Fig. 2. Longitudinal section through genital region, t, testes ; vd, vasa deferentia ; o, ovaries ; 

od, oviducts ; sp, intersegmental septa. 
Fig. 3. Longitudinal section through oviduct, ch, chitinous membrane lining orifice of duct. The 

cells here are not ciliated. 
Fig. 4. Spermatheca as seen on a dissection of the worm, sp, spermatozoa in its interior. 
Fig. 5. Transverse section of spermatheca. b, clumps of granules secreted by cells, and serving 

to agglutinate together the spermatozoa ; n, nuclei ; a, muscular sheath. 
Fig. 6. Setae, a, of dorsal ; b, of ventral pair, drawn to scale. 
Fig. 7. View of four segments in the posterior region of the body, to illustrate the vascular trunks. 

d, dorsal vessel ; v, ventral vessel ; n, nerve cord. 
Fig. 8. Transverse section of one of transverse vessels of anterior segments, e, epithelium lining 

the vessel. 
Fig. 9. Longitudinal section of one of the transverse vessels of anterior segments, m, muscular coat. 
Fig. 10. One of transverse vessels of third segment, p, peritoneal cells forming a thick coating 

round the vessel. 
Fig. 11. Section through 14th segment, a, ripe ova ; n, their nuclei ; sp, intersegmental septum ; 

os, wall of egg-sac. 
Fig. 12. Section through body-wall to illustrate proportionate thickness of e, epidermis; m', trans- 
verse muscular layer ; m, longitudinal muscular layer. 



FE.BEDDARD ON PHREORYCTES 









i 






n 









- 




J%. £. 



if 

: 






I 

Sit < 




Trans. Roy. Soc. Edm r Vol. XXXV. 






^ 






Fig. 5. 




ou\ 



Fig. 6, 



■ , : 









• 



J, 






/v<7. 8. 



a* 

*1 

■, 
(•m 



i 




!.Beddard del 



F Kuth.LiOiTEdm 1 



( 641 ) 



XVII. — On the Placentation of Halicore Dugong. By Sir Wm. Turner, M.B., LL.D., 
D.C.L., F.R.SS.L. and E., Professor of Anatomy in the University of Edin- 
burgh. (Plates L, II., III.) 

(Read 1st July 1889.) 

Comparative anatomists have long desired detailed information on the placentation 
of the Sirenia. It is true that in 1878 Dr Paul Harting presented to the University of 
Utrecht a graduation thesis,* in which he described the foetal membranes and foetus of a 
Dugong, which had been acquired a short time before for the Zoological Museum of that 
University. The specimen had been preserved for many years in spirit of wine, and had 
apparently been collected by a surgeon to a merchant ship. The foetus was 2 7 '8 centi- 
metres (11 inches) long. The entire chorion, with the exception of the two poles and 
their immediate vicinity, was covered by densely packed, short and but little branched 
villi, and Dr Harting came to the conclusion that the placenta of the Dugong was 
diffuse and non-deciduate. 

Dr Harting's description and drawings, though satisfactory as far as the material at 
his disposal would admit, were incomplete, owing to the absence of the uterus. The small 
size of the foetus, in relation to the magnitude of the adult animal, also made it possible 
that the diffused arrangement of the villi over so large a part of the surface of the 
chorion represented an early stage in the placental formation in this animal, and that 
some modification might arise in a later stage of intra-uterine development. I have, 
accordingly, been very desirous to acquire the gravid uterus of the Dugong, and if possible, 
at a more advanced stage of gestation than in the Utrecht specimen. In my endeavours 
to obtain such a specimen, I was fortunate to enlist the sympathy of my friend and former 
pupil Dr Anderson Stuart, Professor of Anatomy and Physiology in the University 
of Sydney, N.S.W.; who, on hearing that C. W. de Vis, Esq., M.A., the Curator of 
the Queensland Museum, Brisbane, had a gravid uterus of the Dugong in his possession, 
which he had received from the Moreton Bay fishers, wrote to ask if it could be sent to 
me. Mr de Vis with the greatest courtesy acceded to this request, and the specimen 
reached me in good order in the month of May of the present year. I wish to express 
my indebtedness to Mr de Vis for having so generously placed at my disposal a specimen 
of so much rarity and value. 

The gravid uterus had been preserved in spirit sufficiently strong to have kept it in 
good condition, but not so concentrated as to needlessly contract the tissues. With the 
exception of two or three cuts of no great size, it was uninjured, though the Fallopian 
tubes had been divided, and the ovaries cut off in taking it out of the abdomen. 

* Het Ei en de Placenta van Halicore Dugong^ Proefschrift, 18th February 1878. Utrecht. 
VOL. XXXV. PART II. (NO. 17). 5 L 



642 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 



Uterus and Maternal Placenta. 

The uterus consisted of two horns, and contained a single large foetus, which was 
situated in the greatly dilated left cornu (fig. l). This the gravid horn measured along 
its anterior convex surface 4 feet 8 inches in its long diameter, and 3 feet 3 inches in its 
greatest girth. It was curved upon itself, so that the part which reached the corpus uteri, 
where the head of the foetus was situated, approximated to the tubal end of the cornu 
occupied by the tail of the foetus. It was invested in the usual way by peritoneum, which 
membrane passed off from it as a broad ligament. The right cornu contained no foetus, 
and that part which was free and surrounded by the broad ligament of peritoneum was 17 
inches long : at its tubal end it was attenuated, but widened somewhat as it approached 
the gravid left cornu, with the wall of which, as was afterwards recognised when the 
uterus was opened into, its wall became closely attached for an additional length of 
10 inches. 

The gravid horn was opened into by an incision along its convexity, extending from 
its tubal end to the corpus uteri. The gravid horn communicated freely with the corpus 
uteri, the commencement of which was differentiated by a strong fold of mucous mem- 
brane. The mouth of the non-gravid horn was also differentiated by a sharp fold of 
mucous membrane from the corpus uteri, and the extended hand could readily be passed 
into that part of the non-gravid horn already referred to, as closely attached to the wall 
of the gravid cornu. The bottom of this dilated portion opened into the lumen of the 
more attenuated part of the non-gravid horn, which barely admitted the tip of a finger. 
The corpus uteri was 11^ inches long; the hand could readily be passed along it for 
several inches, when it became somewhat constricted. Further back it again dilated, and 
the mucous lining, instead of being almost smooth, as in the part of the corpus uteri 
next the cornua, was elevated into distinct longitudinal folds, the best marked of which 
were 0'5 inch in depth. In the intervals between these folds were some transverse and 
oblique folds much less projecting, which gave to this part of the uterus an appearance 
something like that of the mucous lining of a cervix uteri. This part of the uterus 
contained an inspissated secretion. The posterior end of the cervix-like corpus uteri 
communicated by an os with the vagina ; the inferior lip of the os projected more than 
the superior ; and the os itself admitted two fingers. The vagina was 10 inches long and 
5^ inches broad. It was lined by a mucous membrane which was longitudinally folded 
near the os, but further back the folds were oblique and transverse. The urinary bladder 
and urethra were attached to the inferior wall of the vagina. The contracted bladder was 
only 3 inches broad by 2\ inches high, and the pair of ureters opened into its vaginal 
surface. 

The wall of the gravid cornu was much thinner than that of the non-gravid horn, 
in the latter of which the muscular coat greatly exceeded in thickness the corresponding 
tunic in the gravid horn. The mucous lining of the gravid horn had especial interest in 



SIR WM. TURNER ON THE PLACENTATTON OF HALIOORE DTJGONG. 643 

connection with the formation of the placenta, and to the characters of this I shall now 
direct attention. 

The larger part of the mucous membrane lining the gravid horn was smooth on the 
surface, and with no trace of crypts or depressions visible to the naked eye. The surface 
was indeed, in a large part of that area which corresponds to the back and sides of the 
foetus, so smooth as to be almost polished, and with no appearance of succulency, which 
may perhaps in part be due to the hardening action of the spirit in which the uterus had 
been preserved. Near the mouth of the left tuba, which opened into the horn under 
cover of a crescentic fold of mucous membrane by an orifice large enough to admit a 
small surgical probe, the mucous membrane was thrown into folds, and without crypts, 
and some transverse folds were seen in the mucous lining placed opposite the belly of the 
foetus. A broad zone of mucous membrane situated towards the left end of the gravid 
horn presented very different characters, and constituted the maternal placenta. It was 
differentiated from the smooth mucous membrane on each side by a sharp, sinuous margin, 
the left border of which was nearly 8 inches from the opening of the left Fallopian tube, 
whilst the right border was between 2 and 3 feet from the corpus uteri. The zone was 
nearly 1 foot in breadth in the anterior convex part of the horn, and about one-half that 
breadth at its concavity. It was thick and succulent, more especially along the concavity 
of the cornu, where it was elevated into folds, and it was much more vascular than the 
smooth area of the mucous membrane on each side of it. It corresponded in form and 
relative position to a similar zonary band on the chorion. The two zones had indeed 
been closely adapted to each other, but in the handling and shaking to which the uterus 
had been subjected in its removal from the animal and its long journey before it reached 
me, they had become detached. When I opened into the cornu therefore the zone in 
the uterine wall was readily drawn away from the chorionic zone. 

I then proceeded to examine the mucous membrane of the uterus in the smooth 
region beyond the margins of the zonary placenta. With a simple lens the free surface 
was seen to be very faintly wrinkled, and at distant intervals there were faint depressions 
such as might be made from the pressure of the head of a minute pin. The mucous 
membrane was easily removed from the muscular coat by tearing through the submucous 
areolar tissue. When flakes of this membrane were dissected off and placed on a glass 
slide and soaked in glycerine, they became sufficiently translucent to enable one to see, 
with the use of low powers of the compound microscope, the arrangement of the tubular 
glands of the uterus. The mouths of the glands were observed to open at intervals on 
the surface of the mucous membrane, and the direction of the opening was invariably 
oblique (fig. 5). From its mouth the gland tube could be traced obliquely through the 
mucous membrane into the submucous coat. As a rule, the tube was dilated close to 
the mouth, it then diminished in calibre and became cylindriform, and ran for some 
distance as a straight unbranching tube. The tube then bifurcated, and became some- 
what tortuous ; the branches of bifurcation sometimes but not always branched, and 
each terminal branch ended with a closed rounded end. The distinctness of the gland 



644 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DTTGONG. 

tubes varied according to the condition of their epithelial contents. When the cells 
were granular, they and the tubes were more distinct than when the cells had a clearer 
protoplasm. In the field included by a one-inch objective, it was quite exceptional to 
see more than the mouth of a single gland. The glands were relatively of consider- 
able length, from ^ to ^ (7 mm. to 13 mm.) inch, and occupied an area corresponding to 
tln-ee or even four times the diameter of the field, covered by a one-inch objective, and 
as in their course they passed very obliquely from the gland mouth, the mucous 
membrane required to be stripped off in flakes of half an inch in diameter, in order to 
obtain a view of the whole length of a gland. Vertical sections made through the thick- 
ness of the mucous membrane also threw additional light upon the arrangement, and 
enabled one to see the form and characters of the glandular epithelium. In these 
sections the glands were divided sometimes longitudinally, at others obliquely, at others 
transversely, and the segments occupied different planes in the vertical diameter of the 
section; but as the glands were not numerous the segments seen in any given section 
were few, and were separated from each other by relatively broad areas of inter- 
glandular connective tissue (fig. 6). 

As the histological characters of the mucous membrane had been well preserved, the 
form and arrangement of the glandular epithelium could be seen without difficulty. In 
the greater part of the gland-tube the cells were not sufficiently elongated to be called 
columnar epithelium, but were for the most part cubical in form, though some might be 
called polygonal (fig. 8). They were arranged in a single row, and were in part attached 
to the wall of the gland-tube, though in some sections they had become detached from 
the wall and formed clusters of cells lying loose in the lumen. The cells were nucleated, 
and the protoplasm as a rule was granular and opaque. The dilated part of the gland in 
proximity to the mouth was, however, lined by more elongated cells, which could very 
properly be called columnar epithelium. At the gland-mouth itself the cells were not 
distinct, but seemed to be broken down into a formless debris (fig. 7). 

The interglandular connective tissue was also examined. The bundles of this tissue 
were for the most part arranged parallel to the plane of the free surface of the mucous 
membrane, and the membrane had consequently a tendency to split into layers running 
in the same direction. This tissue was loaded with corpuscles, especially in its more 
superficial layers, and, as I have seen in the mucous membrane of many other gravid uteri 
which I have examined, the corpuscles varied in appearance. A large number, and those 
more particularly in the superficial layers, had the spherical form, the size and the 
nucleated granular protoplasm of leucocytes; and it was not unusual to find them 
arranged in large numbers in rows situated between the bundles of connective tissue 
(fig. 7). From their abundance there can, I think, be no question that they have some 
definite functional import. Other corpuscles, more ovoid in form, had a similar kind of 
protoplasm. But in addition, corpuscles, such as one is familiar with in such rapidly 
growing connective tissue as the mucosa of gravid uteri, were seen in large numbers. 
They were most distinct in the deeper layers of the mucosa, where the leucocytes were 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 645 

less abundant, and they had the appearance of fusiform, caudate, or stellate cells. 
Divided blood-vessels were also not unfrequently met with ; sometimes they were empty 
and contracted, at other times the lumen was occupied with red blood corpuscles ; the 
former were presumably arteries, the latter veins. 

As regards the relations of the glands to the interglandular connective tissue, the 
gland wall was usually separated from the surrounding tissue by an interspace (fig. 7), 
which may, perhaps, be a lymph channel. The connective tissue bounding such an 
interspace externally was usually a well-defined bundle arranged in a definite manner 
around the gland-tube. 

I then proceeded to examine the placental zone of mucous membrane from which the 
chorion had been detached. In the first place, I used only a simple lens, and with it went 
carefully over the free surface of the zone, which had a spongy character ; for it was 
perforated by minute orifices, many thousands in number. They were circular, sub- 
circular, or polygonal in outline, and were closely crowded together, being separated from 
each other by delicate bands of mucous membrane (fig. 10). These openings communicated 
with crypt-like depressions in the maternal placenta in which the chorionic villi had been 
lodged. The form and arrangement of these crypts was determined when vertical 
sections through the placental zone were made and examined. Interspersed at intervals 
amidst these minute orifices were somewhat larger openings distinctly visible to the naked 
eye, and with the mouth opening obliquely on the free surface of the placental zone (fig. 
10a). In many of these larger openings the end of a longer chorionic villus was inserted, 
and as the base of the villus had torn away from the chorion in the act of separation of the 
fcetal placenta, these longer villi were left attached to the maternal part of the placenta. 
A careful survey of the whole surface of the zone showed only three patches devoid of 
crypts, the largest of which was If inch in length by 1 inch in breadth, whilst the 
smallest was 9 mm. by 5 mm., and it is possible that these may have been due to the 
abrasion of the crypts from the surface. 

The crypt-like or spongy layer of the maternal placenta varied in thickness from one- 
third of a millimetre to nearly one millimetre. When vertical sections were made through 
it, and examined with a magnification of 60 diameters, the crypts were seen to pass from 
their orifices on the free surface, usually with a considerable obliquity either as far as or 
almost as far as the attached surface of the crypt layer. As a rule, they were elongated 
and tubular, and conformed to the shape of the chorionic villi which they had originally 
contained. Opening into them were smaller secondary crypts, which were cut across in 
the sections, so that the spongy character was not limited to the surface, but prevailed 
throughout the entire thickness of the crypt-layer (figs. 11, 12). The crypts and their 
subdivisions were separated from each other by slender and delicate bands of the mucous 
membrane. The crypts were lined by an epithelium, the cells of which were in part in 
apposition with their walls, and in part had become detached, and were lying loose in their 
cavities. Under higher powers, it was seen that the cells varied in shape ; some were 
longer in one diameter than in the opposite, though the elongation was not so great as to 



646 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 

lead one to call them columnar ; the majority were, however, irregular in shape, and with 
angular outlines, so that they may appropriately be called polygonal cells. The cells were 
nucleated, and as a rule the protoplasm was granular ; they were stained a pinkish tint 
with picro-carmine. It was obvious, therefore, that the crypts possessed a distinct 
epithelial lining. Occasionally I obtained in a section a view of one of the larger crypts, 
to which I have referred in my description of the surface of the maternal placenta. It 
was seen to pass very obliquely and considerably deeper into the mucosa than was the 
case with the smaller crypts, for not only was it wider than they, but it was also three or 
four times longer, and extended obliquely and subjacent to the more superficial crypt- 
layer. In some sections the largely dilated bulbous end of one of the longer villi was 
seen to occupy the interior of a deeper crypt (fig. 11). These deeper crypts also 
possessed small secondary crypts branching off from them, and both were lined by an 
epithelium similar to that found in the shorter crypts already described. It is possible 
that these deeper crypts were the dilated mouths of the uterine glands opening on the 
free surface of the maternal placenta, but it was difficult to trace a gland- tube opening 
directly into any one of them, though in one section (fig. 12) the gland came close up 
to the deep crypt, and seemed as if it might have opened into it. If such were their 
nature, then a few of the chorionic villi, somewhat larger in size than the rest, occupied 
with their free ends the dilated mouths of the gland-tubes. The great bulk of the villi 
had been, however, lodged in the thousands of smaller crypts already referred to, which 
were not the mouths of glands, and indeed were quite independent of them, for they were 
situated in the interglandular portions of the mucous membrane, delicate bands of which 
formed their walls. 

Subjacent to the crypt-layer of the mucous membrane of the placental area was the 
glandular-layer, and in it short segments of the gland-tubes could be seen at irregular 
intervals and at different depths (fig. 11). As in the corresponding sections through the 
non-placental area of the mucous membrane, the tubes were divided either longitudinally, 
obliquely, or transversely. The glands were lined by a cubical epithelium, which was 
either attached to the wall or was lying loose in the lumen. In some sections the gland- 
tubes were seen approaching and indeed lying immediately subjacent to the crypt-layer, 
and in the section above referred to a gland appeared as if it might have opened into 
one of the deeper crypts (fig. 12). 

The connective tissue in the gland-layer contained a large number of corpuscles, 
partly leucocytes and partly those belonging to the connective tissue itself, similar to 
those I have already described in the non-placental part of the mucous membrane. 

An attempt was made to inject the finer vessels of the uterine mucous membrane 
through the uterine veins and arteries, but without success. Although the injection 
passed into the larger trunks, it did not penetrate the smaller arteries and veins, or the 
capillaries, which was doubtless due to the constringing action of the spirit in which the 
uterus had so long been immersed. There can, however, be no question that capillaries 
had ramified freely in the walls of the crypts, and had formed networks similar to those 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DTJGONG. 647 

which. I have injected, and figured in so many of the other specimens of placentae which 
I have described.* 

Attention was then paid to the appearance of the mucous membrane lining the non- 
fecundated cornu. The dilated commencement of this cornu was lined by a mucous 
membrane similar in appearance to the non-placental mucosa in the gravid horn. In 
the non-dilated part of the non-fecundated cornu, the mucous lining was partly smooth, 
and partly elevated into slender, low, longitudinal folds. Vertical sections were then 
made through the mucous lining of the non-expanded horn, and examined micro- 
scopically. It was at once seen that the mucosa was much thinner than in both the 
placental and non-placental areas of the gravid horn. The glands were cut across in 
the sections, but the portions of gland-tube thus divided were much more closely set 
together than in the fecundated cornu, and the interglandular connective tissue was much 
smaller in quantity. These differences will at once be recognised by comparing fig. 6, from 
the fecundated cornu, with fig. 9, from the non-fecundated cornu, which are drawn to the 
same scale. The glands in the fecundated cornu were a little larger than those in the non- 
gravid horn, but they resembled each other in the characters of their epithelial contents. 
Whilst in both regions leucocytes, and connective tissue corpuscles were present in the 
interglandular tissue, it seemed as if in the non-fecundated cornu the leucocytes were 
relatively much less numerous. From a comparison of the membranes in the two cornua 
it is obvious that the great expansion of the mucosa in the gravid horn is very little due 
to dilatation of the glands, but is almost entirely the result of an extraordinary growth 
of the interglandular connective tissue, which has added both to its thickness and super- 
ficial area ; so as to accommodate it to the enormous expansion of the horn as a chamber 
for the lodgment of the foetus, of its membranes and the fluids of the amnion and allantois. 
The great growth of the interglandular connective tissue in the mucous lining of the 
gravid horn in the Dugong, is in accordance with the observations which I have pre- 
viously made and recorded on the placenta in many other mammals. 

Fcetal Membranes and Foetus. 

The chorion was an elongated sac, bent upon itself, and corresponding in form to the 
gravid horn in which it was contained. Its left pole was in proximity to the left tuba, 
its right pole occupied the horn where it communicated with the corpus uteri. Owing 
to the head of the foetus being directed to the right, this pole of the chorion was more 
capacious than the left pole near which the tail of the foetus was situated. The chorion 
was carefully examined to see if it possessed a diverticular prolongation which had 
occupied the dilated part of the right non-gravid horn, but none was seen. 

The outer surface of the chorion was smooth and non- villous in much the larger part 
of its area. The villous part was limited to a zonary band, which passed around the 

* See amongst others the Placenta of the Lemurs, in Trans. Roy. Soc. Lond., vol. 166, 1876 : that of Orca 
gladiator, in Trans. Roy. Soc. Edin., voL xxvi., 1871, and of the Grey Seal, Fox, and Cat in the same, vol. xxvii., 1875. 



648 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 

chorion, considerably nearer to the left than the right pole, so that it was excentric in 
position (fig. 2). The greatest breadth of the zone was on the anterior convexity of the 
chorion, where it measured 11^ inches (292 mm.); in the concavity it was only about half 
that breadth, whilst at a line about midway between it was 9^ inches (241 mm.). In the 
concavity of the chorion, the zone was thicker and folded, whilst on the convexity it was 
thinner and smoother, but the whole surface was flocculent from the presence of multitudes 
of closely compacted villi. The margins of the zone were differentiated from the non- 
villous chorion by a sharp sinuous line. The left border was about one foot from the 
left pole of the chorion, whilst the right border was about 2-^ feet from the right pole,* 
so that if the chorion be supposed to be equally divided into a right and a left half, the 
whole of the zone was situated in the left half. The chorionic zone formed the foetal 
part of the placenta, and it was found to be detached from the maternal placenta, when 
the uterus was opened, as has already been stated. 

I then lifted out of the gravid horn the chorion with its contents, and made a cut 
through that membrane opposite to the spinal column of the foetus from the cephalic 
pole to the right border of the placental zone. A large sac was opened into, which proved 
to be the allantois. This sac was very capacious, and completely surrounded the amnion 
from its cephalic pole to the right placental border of the chorion. The outer wall of the 
sac, smooth, and free on its inner surface, was formed by that part of the allantois which 
constituted the endochorion, and which extended from the cephalic pole of the chorion 
to the right border of the placenta. The inner wall of the sac consisted of the allantois 
which enveloped the bag of the amnion. Immediately subjacent to the right margin of 
the placenta, where it was opposite the back and in part the sides of the foetus, the 
endochorion was reflected on to the outer surface of the amnion, and became continuous 
with the allantois covering that membrane. The sac of the allantois was consequently 
not continued over the inner surface of the placenta in its whole extent, for throughout 
an ovoid area 8 inches in diameter in one direction, and 5 inches in another, which lay 
opposite the back and sides of the foetus, the amnion was in direct contact with the 
inner surface of the placenta. The arrangement of the allantois in relation to the 
part of the placenta which was opposite the belly of the foetus differed materially from 
that above described, for it was prolonged across the inner surface of that portion of the 
placenta from the right to the left border, and its sac was continued into the caudal part 
of the chorion up to the caudal pole. Immediately to the left of the ovoid area, where 
the amnion was in direct contact with the placenta, the allantoic sac was again inter- 
posed between the placental area of the chorion and the amnion, and was continued into 
the caudal pole of the chorion so as to surround the caudal end of the amnion, as com- 
pletely as its cephalic end. Hence the whole extent of the amnion from the caudal to 
the cephalic end of the foetus, with the exception of the ovoid area in relation to a 
limited part of the inner surface of the placenta, was surrounded by and in contact 

* Owing to the poles of the chorion being rounded and bent each on itself, it is difficult to fix with precision the 
central point of the pole. The above measurements are therefore to be regarded as approximative. 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 649 

with the inner wall of the sac of the allantois, which together with the fluid in that sac 
separated the amnion from the endochorion (see woodcut, p. 660). 

Both the endochorion and the allantois enveloping the amnion were vascular, and 
the vessels were derived from the umbilical cord. The cord was undivided for from 
3 to 4 inches beyond the belly of the foetus, and in its undivided part it was invested by 
the amnion. A transverse section made through it immediately before its division 
showed eight vessels, four arteries and four veins, each of which was surrounded by 
delicate gelatinous tissue. As soon as the cord reached the allantois it divided into 
four branches, each of which contained an artery and a vein. Each branch immedi- 
ately became invested by a fold of the allantois, by means of which it was conveyed 
to the inner surface of the placenta opposite the belly of the foetus. These folds were 
of considerable length and breadth, and were continuous with that layer of the allantois 
which enveloped the amnion. They and the branches of the cord situated along their 
free borders had floated in the fluid of the allantoic sac. The cord was spirally twisted 
before its division, each of its branches also was tortuous, and a similar character also 
was seen in the vessels as they ramified on the inner surface of the placenta. An 
attempt was made to inject the umbilical vessels with gelatine and carmine, which was 
successful as far as the large trunks were concerned. Many of the vessels ramifying 
on the inner surface of the placenta were also filled. The injection had not penetrated 
into the capillaries within the villi, though it was not unusual to see villi stained with 
carmine from the injecting material. 

The vascularity of the chorion was not limited to the placental zone, for a number 
of vessels, both arteries and veins, passed beyond each margin of the placenta as far as 
the corresponding pole of the chorion. To some extent the injection had penetrated 
into these vessels, but even when not injected they could be readily seen to ramify in 
the endochorion lining the non-zonary parts of the chorion. 

The villi in the placental zone varied in length. Comparatively a few were on the 
average T 7 u^ ns mcn (18 mm.) long, but the great majority, many thousands in number, 
were from -j^ths to T ^ths inch (8 to 10 mm.). The stems of the villi were cylindriform 
and filamentous, but those of the longer villi were thicker and tougher than the stems 
of the shorter villi. The shorter villi branched about x^th inch (2*5 mm.) or a little more 
from their origin from the chorion, and gave rise to two, three, four, or more rarely 
five branches, which arose in proximity to each other (fig. 3). They were much smaller 
than the stems from which they sprang, and after a course of about T %ths inch they not 
unfrequently bifurcated, and the branches of bifurcation ended in short terminal buds. 
These shorter villi had been lodged in the shorter crypts which formed the spongy 
surface of the uterine mucosa, and had been drawn out of the crypts before I opened 
the uterus. 

The longer villi were more firmly attached to the maternal zone, so much so indeed 
that they had been torn away from the chorion, and their free ends were implanted in 
the wider and deeper crypts described in my account of the maternal placenta. The free 

VOL. XXXV. PART II. (NO. 17). 5 M 



650 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 

end was dilated into so bulbous an expansion as obviously to interfere with, the ready 
extraction of the villus from the crypt (fig. 11). 

The substance of each villus consisted of a delicate connective tissue such as one is 
accustomed to see in these structures. Many had a definite epithelial investment, the 
cells of which were irregularly polygonal, but others had lost their epithelial coat. 
What, however, was especially noticeable was the large amount of cells, either singly 
or in groups, which were seen floating in the fluid in which the villi were examined after 
they had been teased asunder with needles. These cells consisted in part without doubt 
of the epithelial investment of those villi which had lost their covering probably in the 
process of teasing ; but from the numbers seen I think that many were derived from the 
lining epithelium of the maternal crypts, which having been detached and intermingled 
with the villi, had come away with them when the fcetal placenta separated from the 
maternal. 

Projecting from the inner surface of the endochorion into the allantoic sac were 
numbers of flattened leaf-like bodies, which I shall name the allantoic bodies. They 
were arranged either singly, or in pairs, or sometimes in rows of several situated close 
together. They were confined in their position to the allantois close to the placental 
borders, or to that part of the placenta which was covered by the allantois. They grew, 
therefore, from the most vascular part of the membrane, and were continuous with. the 
coats of the blood-vessels, more especially the veins. These allantoic bodies varied in 
shape and size. Some were elongated and attached by narrow pedicles to the endochorion ; 
others were broader than long, and with broad bases of attachment, and when a row of 
them occurred close together along the course of a particular vessel they sometimes were 
continuous with each other at the base of attachment. The longest of these bodies was 
f of an inch, but ^ an inch (13 mm.) was a more usual size. The base of attachment 
was in many cases f of an inch, but when two or more had fused together the base was 
much more elongated. The smallest allantoic body which I measured was 3 to 4 mm. 
in diameter. In colour they were light brown, like that of the allantoic membrane, 
but this was probably not the natural tint, as the membranes were all stained and 
discoloured. All these bodies were flattened, but this may possibly have been due in 
part to post-mortem compression, after the allantoic and amniotic fluids had been 
evacuated, and the weight of the foetus had pressed against them. 

Some of the allantoic bodies were then removed along with the part of the vessel 
to which each was attached and prepared for microscopic examination, and sections 
were made through them and their blood-vessel from the attached to the free border. 
They were seen to be continuous with the fibrous connective tissue coat of the vein, to 
which they were attached by the base, and in tracing them from the base to the free 
border they became more and more attenuated. They consisted for a part of their 
thickness of wavy bundles of white fibrous connective tissue; but in the axis of the body 
the tissue was modified in appearance. When teased out with needles elongated fibro- 
cells could be isolated, the characters of which were scarcely sufficiently differentiated to 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DTJGONG. 651 

permit one to say whether they were more like embryonic connective tissue, or involun- 
tary muscular fibres, though I am inclined to think the former, for the nuclei in asso- 
ciation with them were more fusiform than rod-like. Small blood-vessels were also seen 
to be cut across in the sections. 

I did not obtain any evidence of the presence of an umbilical vesicle. 

The amnion was a capacious sac for the lodgment of the foetus with its amniotic 
fluid. It extended from the cephalic to the caudal pole of the chorion, and throughout 
its entire extent, except at the ovoid area on the placenta already described, it was 
separated from the chorion and placenta by the inner layer of the allantois and the 
allantoic sac. Its poles, therefore, were free, and could be readily moved to and fro in 
the allantoic cavity. Its outer surface was united to the inner layer of the allantois by 
delicate areolar tissue, which, on being torn through, enabled the two membranes to be 
separated from each other. Its inner surface was smooth and polished, though in places 
blood-stained and with adherent blood-clots, due to two punctured wounds in the uterus 
and foetus. It was particularly noticed that the amniotic covering of the cord, as well as 
the inner surface of the sac, was free from such granular projections as have been seen by 
various observers in considerable numbers studding the free amniotic surface in the 
Cetacea,* and which I have named the amniotic corpuscles. 

The foetus was a male. It occupied the bag of the amnion, being curved on itself, so 
that the tail was bent forward under the hinder part of the body, and the head was bent 
back towards the chest. In consequence of this curvature the tail concealed the penis 
and anus, and the abdomen was marked with deep transverse folds. The pectoral limbs 
were directed backwards and somewhat ventralty on the sides of the trunk. There was no 
dorsal fin. The skin was smooth, and of a dull yellowish-grey tint, lighter on the belly 
than on the sides and back. Scattered, delicate, silky hairs from ^yths to x&ths of an 
inch long, projected through the skin, more numerous on the head and body than on the 
limbs and tail. The mouths of the hair follicles were very distinct, their position being made 
more clear by the integument immediately surrounding each follicle being of a paler hue. 
The hairs and follicles were arranged on the back of the foetus in rows having an antero- 
posterior direction. In the intervals between these follicles a number of much finer spots 
were interspersed. These were apparently the mouths of follicles for smaller and more 
delicate hairs, but the hairs were not projecting from them. From the front of the 
muzzle a moustache, consisting of short, stiff, white hairs, projected, and more delicate 
hairs grew out of the skin of the lower lip. 

The muzzle was very characteristic. It was flattened at the front, and consisted of 
two lateral halves, with a median portion. The lateral halves sloped downwards and 
outwards so as to conceal laterally the lower lips and mouth slit ; but in front the mouth 
slit was seen below the median portion of the muzzle, which was a tongue-like lobe pro- 
jecting from the anterior end of the roof of the mouth. A mesial groove, commencing 

* I may refer to a detailed description of the amniotic corpuscles in my account of the Placentation of Orca 
gladiator, in Trans. Roy. Soc. Edin., 1871, vol. xxvi. 



652 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 



just above the tongue-like lobe, ran upwards for If inch, when it ended in a shallow 
transverse groove not quite \ inch long. On each side of the mesial groove was a lateral 
groove of about the same length, which passed obliquely downwards and outwards to the 
most projecting part of the lip on its own side. The palpebral fissure was small, so that 
only a limited surface of eye-ball was exposed. The auditory meatus was about the size 
of the head of a small pin. The two anterior nares were situated above the muzzle ; they 
were crescentic in shape with the concavity downwards, and were about -j^ths of an inch 
asunder ; each nostril was -^ths of an inch wide and T ^ths of an inch in vertical diameter, 
and had a flap-like piece of skin at the orifice. The pectoral limb was paddle-shaped, and 
with a slight indentation in its posterior end near the ulnar border. The lobes of the 
horizontal tail projected behind the mid point, so that the hinder border was concave, but 
there was no mesial notch. The anal orifice was relatively large. The penis projected 
through a circular hole surrounded by a fold of skin which was 5 inches in front of the 
anus. The umbilical cord projected from the wall of the abdomen 2 inches in front of 
the orifice for the penis. It is difficult to get exact information of the length of the 
adult Dugong, as the statements in the books vary from 8 or 10 feet to 20 feet. On the 
supposition that the adult is only 10 feet long, then this foetus had attained quite one- 
half of the adult length. 



Dimensions of Fcetus. 



Length along curve of back, 
From middle of tail to anus, 
From anus to orifice for penis, 
From orifice of penis to umbilicus, 
From axilla to tip of pectoral limb, 
Greatest breadth of pectoral limb, 
From angle of mouth to eye, 
Width of mouth slit, 
From mouth to nostril, 
From eye to external meatus, 



Feet. Inches. 
5 4 

1 1 

5 

2 
7 

- *U 



d 4 



I am also indebted to Mr de Vis for a much smaller foetus from another uterus, which 
was, however, unaccompanied by any of the membranes. It was a male, and was 
curved on itself like the larger specimen. The surface of the skin was a pale drab 
colour, and was marked; with spots which obviously showed the position of the hair 
follicles, though no hairs were to be seen, neither was any moustache visible. The 
muzzle showed the mesial tongue-like process, which was relatively larger than in the 
older specimen. The muzzle was flattened in front. The eye-balls were relatively large, 
but the palpebral fissure was very small. The anterior nares were small, and the auditory 
meatus was so minute as to be scarcely visible. The tail was bilobed, but the mid point 
projected a trifle further back than the lateral lobes. The pectoral fin was paddle-shaped, 
and with a similar indentation in its border to that described in the older fcetus. There 
was no dorsal fin. This fcetus was only 14 centimetres long (5^ inches). 



SIR WM, TURNER ON THE PLACENTATION OF HALICORE DUGONG. 653 

General Observations on the Placentation of the Dugong. 

I have already stated that a description of the foetal membranes of the Dugong was 
given a few years ago by Dr Harting of Utrecht. His specimen was at a much earlier 
stage of development than mine, for his foetus was only 27*8 cm. long, whereas mine 
measured 163 cm., i.e., 5 feet 4 inches. The chorion presented a very different appear- 
ance in the two specimens. Harting describes it in his example as being, with the 
exception of its poles, covered on its entire surface with villi ; the length of which varied 
from 0'7 to 1"5 mm. They were most dense at the middle of the chorion, where the 
branches of the umbilical vessels reached the chorion. The villi extended to about 
3 cm. from the posterior pole, but had amidst them spaces bare of villi, so that the 
border of the chorion adjoining the naked pole was very sinuous. The villi approached 
the anterior pole to from 4 to 5 cm. The villi were formed of connective tissue covered 
by a layer of round or oval cells, and contained capillary blood-vessels. The posterior 
non- villous pole had probably been in relation to the tubal end of the uterine cornu and 
the anterior non- villous pole to the ostium uteri, but in the absence of the uterus their 
relations could only be surmised. Close to the anterior pole of the chorion was an 
appendage covered by villi, which Harting thought might have been lodged in the 
non-fecundated horn of the uterus. Owing to the distribution of the villi over so large 
a proportion of the surface of the chorion, Harting came to the conclusion that in the 
Dugong the placenta was diffused as in the Cetacea, the Pig, and the Mare. 

My dissection of the gravid uterus in a much more advanced stage of development 
puts the placentation of this animal in quite a new and different light. As has been 
stated in the description, the placenta is zonary, the zone not being situated around the 
equator of the chorion, but somewhat nearer to the caudal end of the embryo ; and the 
zone is as perfect in form as in the Cat or other carnivorous animal. 

It will be necessary, therefore, to consider how the change from the more diffused 
arrangement seen by Harting to the zonary form observed by me was brought about. 
For this purpose I may refer to observations which I made some years ago on the 
placenta of the common Cat at different stages of development.* In an early stage the 
whole chorion, except an area yxjth inch in diameter at each pole, was covered with villi. 
In a more advanced stage, the non- villous area was somewhat larger at each of the poles, 
though much the greater part of the chorion was still villous, and the placenta had 
assumed the form of a broad zone (fig. 4). In the latter half of gestation the villi 
were limited to a comparatively narrow zone surrounding the equator of the chorion, 
and a much larger area was smooth and free from villi than that which was occupied by 
the placenta. 

Corresponding changes in the proportion of the villous to the non- villous chorion 
take place also, I believe, in the development of the placenta in the Dugong. At first 
the chorion is in all probability entirely covered with villi as is the case indeed at an 

* See for further details my Lectures on the Comparative Anatomy of the Placenta, p. 72, Edinburgh, 1876. 



654 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 

early stage of development of the human ovum. Then as it becomes elongated the villi 
disappear from the poles, so that a definite non- villous area appears at each pole, and with 
the continued growth of the ovum the proportion of the bare spaces to the villous chorion 
increases. This is apparently the stage of development at which Haeting's specimen 
had arrived, and it will be observed that he saw not only the naked poles, but the 
presence of bare patches amidst the villi, which obviously marked the beginning of a 
further disappearance of these structures. In my specimen the placenta had undoubtedly 
reached its completed stage of development as a definite zonary band. 

Another point of difference in the chorion at these two stages of development must 
also be referred to. In Dr Haeting's example there was an appendage covered with 
villi apparently extending into the non-fecundated cornu. In my specimen no such 
prolongation was present, and one can see why in the more advanced stage of develop- 
ment such an appendage was unnecessary. So long as the villi were diffused over a large 
area of the chorion, there must have been a corresponding diffusion of the crypts in the 
uterine mucous membrane for their reception. Consequently at this early stage of 
development, as much of the mucous lining of the non-fecundated cornu as corresponded 
with the villous surface of the appendage of the chorion w r ould contribute to the forma- 
tion of the placenta. But with the gradual atrophy of the villi from the poles and from 
a large part of the surface of the chorion, the corresponding crypts in the mucous 
membrane would also disappear ; the appendage would become functionally inactive, 
and would either atrophy, or become, by the great dilatation of the chorion in the gravid 
horn, incorporated with or expanded into it. 

Another and very important conclusion to which Dr Haeting arrived was that the 
placenta in the Dugong was indeciduate. This indeed was almost a necessary corollary 
from its diffused structure, for so far as we at present know every diffused placenta is 
non-deciduate. But the recognition of the zonary form in the more advanced stage of 
placental development of the Dugong has introduced another element into the discussion 
of this subject. Our knowledge of the zonary placenta has hitherto been confined to its 
presence in the Carnivora proper, the Pinnipedia, Hyrax, and the Elephant. In all of 
these animals the placenta is deciduate, i.e., when shed in the course of parturition the 
chorion carries away with it distinctly recognisable portions of the vascular mucous coat 
of the uterus. On a priori grounds we should expect to find that the Dugong, with its 
zonary placenta, was also deciduate. I examined the placenta, therefore, with great 
interest with the view to determine this question. 

The relative shortness of the great majority of the chorionic villi, and the corre- 
sponding shallowness of the uterine crypts in which these villi had been lodged, together 
with the cylindriform shape both of villi and crypts and the paucity of their branches, 
had permitted the separation of the foetal and maternal divisions of the placenta from 
each other before the uterus was opened into. The separation had apparently taken 
place as readily as in the Mare or Cetacean, in which animals, though the villi and crypts 
are short, yet the villi have numerous tuft-like branches, and the crypts are subdivided 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 655 

into corresponding recesses. As the Mare and Cetacean are typical examples of mammals 
with a non-deciduate placenta, the Dugong also in the general scheme of its placenta 
without doubt resembles them. But my description has shown that, in addition to the 
multitudes of short villi and shallow crypts, the Dugong also possessed a small proportion 
of longer villi, which were implanted in longer, wider, and more deeply seated crypts, 
passing for some distance in an oblique direction subjacent to the layer of short crypts. 
From the fact that in the separation of the maternal from the foetal part of the placenta 
which had occurred before I opened the uterus; these longer villi had torn away from the 
chorion, and had remained implanted in the more deeply seated crypts, it is evident that 
the villi were more difficult to draw out of the crypts, and it is possible, therefore, that 
in the normal act of parturition, when these villi separate along with the rest of the 
foetal placenta, they may drag away the vascular walls of the maternal crypts in which 
their bulbous ends are implanted. Should this be the case, then the placenta of the 
Dugong would in a limited sense be deciduous, though in far the greater part of its area 
it would, as already stated, be non-deciduate. If I am right in this supposition of the 
shedding of the vascular walls of those maternal crypts in which the longer villi are 
inserted, whilst the shorter crypts retain their attachment to the uterus, then the 
placenta, whilst in the main non-deciduate, would furnish an example of the commence- 
ment of the deciduate type, and would illustrate the transition from a non-deciduate 
placenta to the more perfect deciduate type seen in the Carnivora and other zono- 
placental mammals. Should, however, these larger villi separate from the maternal 
placenta without taking with them the vascular walls of their crypts, which I think is 
the more likely, the placenta would be throughout non-deciduate. The placenta of the 
Dugong would then furnish us with a new type of placenta, one which is both zonary 
and non-decicluate. The presence of these two characters, hitherto unknown as existing 
in combination in any placenta previously examined, teaches us that the similarity in form 
which the Dugong's placenta bears to that of the Carnivora and Seals does not necessarily 
imply a correspondence in the intimate relations of the structural elements. We cannot 
therefore predicate whether a placenta is deciduate or non-deciduate from the study 
merely of its external form. 

In the discussion of this question I have used the term deciduate in the sense in 
which it is usually employed to express the shedding of the vascular part of the maternal 
placenta during parturition. In a communication made to this Society a Dumber of 
years ago,* I showed that both in the Sheep and Cow the epithelial lining of the uterine 
crypts was to some extent shed along with the tufts of the foetal villi. It is not unlikely 
that a partial shedding of this epithelium may also take place in the Dugong, for, as I 
have already explained (pp. 645, 650), not only were the epithelial cells in many cases 
loose in the crypts, but others had become so far detached as to be intermingled with 
the villi of the chorion. 

* Proc. Roy. Soc. Edin., May 1875; and Lectures on the Comparative Anatomy of the Placenta, p. 108, Edinburgh, 
1876. 



656 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 

A point of much interest in the study of the placenta in any mammal is the 
relation between the crypts in which the villi are lodged and the glands of the uterine 
mucous membrane. On the supposition which w T as at one time entertained that the 
crypts were the dilated mouths of the uterine glands, there would require to be as many 
glands as there are villi in order that they may be mutually accommodated. The 
inadequacy of this hypothesis was, however, pointed out some years ago both by Signor 
Ekcolani and myself, from the study of the gravid uterus in the Pig and Mare, in 
which animals we showed that the glands opened on definite areas of the mucous 
membrane opposite portions of the chorion where there were no villi, and that the 
uterine crypts were interglandular in position.* Subsequently, in an account of the 
placentation of the Lemurs, I described a yet more striking example of the inter- 
glandular position of the crypts, t In the Cetacea again, as illustrated by Orca 
gladiator,\ I found that though a small proportion of the crypts was associated with 
glands, the great majority had no such relation, and were interglandular in position. 
The Dugong also exemplifies the accuracy of this view. If in it all the villi had been 
lodged in the mouths of glands, then thousands of glands would have been needed for 
their reception ; whereas, as we have already seen, the glands were sparing in number 
in the placental area. The only crypts which seemed indeed in the Dugong as if they 
might be associated with glands were those wider and more deeply situated ones in which 
the longer villi were lodged. But the shorter crypts, which formed the great majority of 
the recesses for the villi, were interglandular in the Dugong, and were produced without 
doubt by an uprising and active growth of the interglandular tissue of the mucous 
membrane around the villi. The villus and the wall of its crypt were contemporaneous 
in their growth, and the depth of the crypt corresponded to the length of the villus. 
The epithelial lining of the crypts is also an important factor in their construction, and 
its discovery in the Dugong brings its placenta in this particular into harmony with 
those of the other mammals that I have described. § 

One or two other points of comparison with the specimen described by Dr Harting 
may now briefly be referred to. In both the allantois was a capacious sac, and 
extended to opposite pt>les of the chorion, and conveyed blood-vessels in its endo- 
chorionic layer ; and the arrangement of the larger trunks of the umbilical vessels was 
also similar in the two specimens. Both also possessed allantoic bodies, which differed 
somewhat in colour, form, and structure in the younger and more advanced specimens. 
In Harting's young example they were yellowish- white, round or ovoid, and made up 
of areolar tissue arranged so as to bound a large number of small cavities, the walls of 
which were constituted of interlacing elastic fibres united into bundles, and between 

* Lectures on the Comparative Anatomy of the Placenta, Edinburgh, 1876. 

+ Trans. Roy. Soc. Lond., vol. 166, 1876. 

X Trans. Hoy. Soc. Edin., vol. xxvi., 1871. 

§ For a discussion on the physiological relations of the epithelial lining of the crypts, I may refer to the section 
headed "Physiological Remarks," in my Lectures on the Comparative Anatomy of the Placenta, p. 114, Edinburgh, 
1876. 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 657 

which small fusiform cells were seen. Within the areolae some cells resembling leuco- 
cytes were found. These bodies were invariably attached to the coats of the blood- 
vessels, almost always to the veins, and Harttng regards them as diverticula from the 
vascular wall. In my older example they were also closely attached to the coats of the 
umbilical vessels of the endochorion. They were light brown in colour, flattened or 
plate-like in form, and not subdivided internally into areola?. They were, however, 
obviously similar structures in the two specimens, and the differences at the two ages 
were doubtless due to the one being structurally more advanced than the other. 

These allantoic bodies in the Dugong are probably homologous with the well-known 
" hippomanes" of the Mare, which, though usually found floating in the fluid contents of 
the allantois, yet arise, as I pointed out some years ago,* in the gelatinous tissue which 
connects the chorion with the endochorion. They also correspond with the shot-like white 
spherical bodies which Professor Rolleston t and I have described in the corresponding 
gelatinous tissue in the Pig ; J which bodies are also attached to the walls of the blood- 
vessels. Sir Richard Owen has shown the presence of similarly situated bodies in the 
fcetal membranes of the Elephant § at about the middle of gestation, most numerous 
near the placenta, and developed in connection with the coats of the umbilical vessels. 
From a microscopic examination, to which I subjected a specimen some years ago, I 
found || that it consisted of a tough fibrous tissue in which nuclear-looking particles 
were imbedded. Dr Chapman of Philadelphia has since described IF the membranes of 
an Elephant delivered at the full term, and has figured these bodies attached to the 
umbilical vessels, both veins and arteries, on and near the placenta. He considers 
them to project towards the amnion, as the allantoic sac had disappeared in the placenta 
at the full term. He states that they are fibrous in structure, with some interfibrous 
granular matters. It is difficult to give a satisfactory explanation of the function of 
the allantoic bodies. 

Neither in Dr Harting's specimen nor in mine were any amniotic corpuscles to be 
seen, nor was there any evidence of the presence of an umbilical vesicle. 

It will now be of interest to compare the placentation of the Dugong, which may be 
taken as a type of the Sirenia, with that of those orders of mammals with which the 
Sirenia have from time to time been regarded as most closely allied. 

By the Messrs Cuvier the Sirenia were grouped with the Cetacea as a suborder, 
Cetacea Herbivora. Although the Sirenia have without doubt some points of corre- 
spondence with the Cetacea in their placental relations, they possess many more features 
of difference. Both orders of mammals have a bicornuate uterus and are uniparous. 

* Lectures on Comp. Anat. of Placenta, p. 26. 
t Trans. Zoolog. Soc, vol. v., 1863. 

X Lect. Comp. Anat. Placenta. C. Hennig has subsequently published a description of these bodies in the Pig 
(Sitz. der Naturf. Ges. zu Leipzig, 1877). 

§ Trans. Roy. Soc. Lond., 1857 ; Anat. Vertebrates, vol. iii. 

|| Lect. Comp. Anat. Placenta, p. 27. 

IT Jour. Acad. Nat. Sci. Philad., vol. viii., 1880. 

VOL. XXXV. PART II. (NO. 17.) 5 N 



658 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 

In the Cetacea the chorion occupies both uterine horns ; in the Dugong it is situated 
in only one horn, at least in the later stage of gestation. In the Cetacea the villi are 
diffused over "both horns of the chorion, except at the poles and opposite the os uteri 
and the uterine crypts are also diffused; in the Dugong in the fully-developed placenta, 
the villi are aggregated into a zone on one side of the equator of the chorion, and there 
is a corresponding limitation of the uterine crypts. In the Cetacea the allantois though 
extensive, and lying in both horns, does not reach the poles of the chorion, as it does in 
the Dugong; and whilst the latter possesses allantoic bodies, there are none in the 
former. In the Cetacea the amnion extends into both horns, reaches beyond the allantois, 
and possesses numerous amniotic corpuscles. In the Dugong the allantois extends 
beyond the amnion, and the latter has no corpuscles. Neither in the Cetacea nor the 
Dugong has an umbilical vesicle been recognised. The Cetacea are undoubtedly non- 
deciduate. The Dugong is also without doubt in the main non-deciduate, though there 
is the possibility of an exception in the case of the longer villi and the walls of the 
deeper crypts. But in this matter I would again point out that in the separation of the 
foetal from the maternal placenta, which had taken place in my specimen before I opened 
the uterus, the longer villi had detached themselves from the chorion and remained 
imbedded in the uterine wall ; whilst on the hypothesis of this part of the placenta being 
deciduate, one would have expected rather that the longer villi would have continued to 
be attached to the chorion, and the walls of the deeper crypts to have torn away from 
the uterus and have enclosed them. There is nothing in the structure of their respective 
placentas to justify the retention of the Cetacea and Sirenia in the same order. 

By some zoologists the Sirenia have been regarded as allied to the Ungulata. 
Amongst ungulate mammals the placenta is not uniform in its arrangement. The Pigs, the 
Equidae, the Camelidee, Tragulidse, the Tapir, Hippopotamus, and probably the Rhinoceros, 
furnish typical illustrations of the diffused variety of placenta ; the Ruminants generally 
exemplify the cotyledonary form ; whilst the Giraffe * and Cervus mexicanus,i in addition 
to a number of large and small cotyledons, had patches of villi diffused over portions of 
the chorion intermediate to the cotyledons, so that they combined the characters of both 
types. All these animals agree in possessing a large and persistent allantoic sac. The 
umbilical vesicle, however, either disappears at an early period of utero-gestation, or is 
so small as to be recognised with difficulty. All ungulate mammals, which possess a 
diffused or cotyledonary placenta, are non-deciduate in the sense in which that term is 
usually employed. 

Hyrax, which has certain ungulate affinities, also possesses a zonary placenta, which, 
as I showed some years ago,| corresponds in structure with that of the domestic Cat. 
The sac of the allantois is so large as to reach the opposite poles of the chorion, but the 
umbilical vesicle apparently atrophies at a comparatively early period of utero-gestation. 

* See my Lectures on the Placenta, p. 67. 

+ See my description of the Placenta of the Mexican Deer, in Journal of Anatomy and Physiology, vol. xiii. p. 195, 
1879. 

t Proc. Roy. Soc. Lond., Dec. 16, 1875. 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 659 

The ungulate affinities of the Elephant have been dwelt upon by various naturalists. 
This animal has a well-marked zonary placenta, which Sir Richard Owen stated to be 
in the middle of the chorion, though Dr Chapman in his more advanced specimen found 
it to be on one side of the middle. In addition, the chorion possessed a small vascular 
patch of villi diffused at each pole of the chorion. The deciduate character of this 
placental zone was recognised both by Owen and myself ; and Chapman estimates, from 
the injection of the blood-vessels in his specimen, that at least one-fourth of the shed 
girdle-like placenta consisted of the hypertrophied mucous membrane of the uterus. In 
Owen's specimen the allantois was persistent ; in Chapman's it is said to have disappeared 
as a distinct sac. Owen makes no mention of an umbilical vesicle, and Chapman states 
distinctly that he did not find a trace of it. De Blainville regarded the Sirenia as a 
group of aberrant Elephants, and more recently Dr Murie * has referred to the probable 
affinities between the Sirenia and the Elephant. 

Many years ago Sir Richard Owen t pointed out that the placenta is not an organ 
which guides to the true grouping of the Ungulata, and that the modifications of the 
placenta do not form a true guide to the affinities or classification of the Mammalia 
generally. In more than one memoir J I have also shown that the placenta cannot be 
taken as a dominant organ for purposes of classification; yet, in the study of the 
affinities of animals, its characters require to be considered in conjunction with those 
of the other organs in the body. In this connection, therefore, it is interesting to 
observe that, as regards its form, the placenta both in the Elephant and the Dugong 
is zonary; though they differ in this very important particular, that in the Elephant 
the zonary placenta is deciduate, in the Dugong it is almost entirely, if not entirely, non- 
deciduate. 

The zonary placenta possesses its best known form and structure in the Carnivora 
proper and in the Pinnipedia, in which animals it is usually equatorial in position. In 
both these suborders it is highly deciduate, although the relative amount of mucous 
membrane which separates along with the chorion would seem to vary in different species. § 
All these animals possess a persistent and well-developed allantoic sac, and as far as we 
at present know they have also a persistent umbilical vesicle. Except in the mere form 
of the placenta, the Dugong does not seem to have any special affinity with the Carnivora 
proper, or with the Pinnipedia. 

J. F. Brandt in his great memoir, " Symbolao Sirenologicse," [| compares the Sirenia in 
their structure and habits with the Cetacea, Pachydermata, Phocacese, and Zeuglodon. 
Although he considers that in some respects they are more closely allied with the 
phytophagous Pachyderms than with any other order of mammals, yet from the evidence 

* " Form and Structure of the Manatee," Trans. Zool. Soc. Lond., vol. viii., 1870. 
t Memoir on the "Placenta of the Elephant" (Trans. Boy. Soc. Lond., 1857). 

X Memoir on the " Placentation of the Sloths" (Trans. Boy. Soc. Edin., 1873, vol. xxvii.) ; and on the " Placentation 
of the Lemurs," in Trans. Boy. Soc. Lond., 1876, vol. clxvi. 

§ See my Lectures on the Gomp. Anat. of the Placenta, p. 111. 

|| M4m. de VAcad. Imp. des Sciences de St Petersb., 7th series, fasc. ii. and iii., cxii., 1869. 



000 



SIR \VM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 



at his disposal he regards them as forming a distinct order. He had no knowledge of 
the form and structure of their placenta, but from the description which I have given 
of that organ in the Dugong, it is clear that its type of structure is so peculiar as to 
strengthen the view that the Sirenia have an ordinal value. 

From the study of the placenta of the Dugong it is obvious that in future it will be 
necessary to arrange the mammals which possess a zonary placenta into two groups : 



ZONO-PLACENTALIA. 

a. Non-deciduata. — Dugong either whole or in great part ; probably also the 
Manatee. 

b. Deciduata. — Carnivora, Pinnipedia, Elephant, Hyrax. 



WOODCUT. 

The subjoined woodcut is a diagrammatic representation of the arrangement of the foetal 
membranes, and their relation to the embryo Dugong; p. 648. The embryo has been drawn in the 
amniotic sac by Mr Gustav Mann. 




Ch, the chorion, with n the zonary placenta. Al, the sac of the allantois; Al', the outer wall of 
the sac or endochorion; AV, the inner wall of the sac. Am, the sac of the amnion, the 
membranous wall of which is represented by the Jotted line. 



SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 661 

EXPLANATION OF PLATES I, II., III. 

Plate I. 

Fig. 1. A view of the gravid uterus of Halicore Bugong, greatly reduced : /, the fecundated left 
uterine cornu ; nf, the non-fecundated right cornu ; v, the vagina ; uu, the two ureters 
opening into the bladder ; p. 642. From nature, by my pupil, Mr Harry G. Melville. 

Fig. 2. The chorion extracted from the uterus of the same animal, showing the zonary placenta 
and the umbilical vessels ramifying in the smooth parts of the chorion : h, opposite the 
head of the foetus ; a, opposite the anterior extremity ; t, opposite the tail ; u, opposite 
the umbilical cord, greatly reduced ; p. 648. Figs. 2 and 4 were drawn from nature by 
Mr Michael Scott. 

Fig. 3. One of the shorter villi of the chorion detached from its neighbours, and with its branches 
drawn asunder; x 13; p. 649. From a drawing by the senior demonstrator of anatomy, 
David Hepburn, M.B. 

Fig. 4. Chorion of a Oat in an early stage of gestation, to show how extensively the villi are diffused 
over the surface, also the non-villous character of the poles ; natural size ; p. 653. 
From a specimen which I have placed in the series of placentas in the Anatomical 
Museum of the University of Edinburgh. 

Plate II. 

Fig. 5. Surface view of the mucous membrane of the left fecundated cornu of the Dugong in the non- 
zonary area. The uterine glands may be seen ramifying in the mucous membrane, and 
the dilated mouths open obliquely on the surface; x 30; p. 643. The figures in this 
Plate, and fig. 10 in Plate III., were drawn from nature, by my pupil, Mr Gustav Mann. 

Fig. 6. Vertical section through the mucous membrane of the same cornu in the non-zonary area. 
Sections through the gland-tubes occur at different depths, and separated by relatively 
wide intervals occupied by interglandular connective tissue : m, the mouth of a gland 
opening on the surface ; v, a transversely divided blood-vessel ; mc, muscular coat; X 80; 
p. 644. 

Fig. 7. Vertical section through the mouth of the gland, m, shown in fig. 6 and the adjoining part 
of the mucous membrane; x 400. The oblique direction of the gland-mouth and the 
appearance of the columnar epithelium cells which occupy the parts of the gland-tubes 
that lie near the surface of the mucous membrane are shown. The interglandular 
connective tissue contains numerous leucocytes, I, as well as connective tissue corpuscles ; 
v, a transversely divided blood-vessel ; p. 644. 

Fig. 8. Vertical section through the closed end of a gland in the deeper part of the mucous 
membrane ; x 450. The epithelial lining of the gland consists of cubical cells ; p. 644. 

Fig. 9. Vertical section through the mucous membrane of the non-fecundated right cornu of the 
same uterus ; x 80. The gland-tubes are smaller than in the fecundated horn, and as 
the interglandular connective tissue is much less abundant, the tubes are much closer 
together. Compare this figure with fig. 6. v, blood-vessels divided ; mc, the muscular 
coat; p. 647. 

VOL. XXXV. PART II. (NO. 17). 5 O 



662 SIR WM. TURNER ON THE PLACENTATION OF HALICORE DUGONG. 



Plate III. 

Fig. 10. Surface view of the mucous membrane of the fecundated cornu in the placental zone. 
The orifices of a number of the smaller crypts are represented, and at a the mouth of 
one of the large crypts ; x 30 ; p. 645. 

Fig. 11. Vertical section through the mucous membrane of the placental zone ; x 60. The figure 
shows the crypt layer, cr, and the glandular layer, gl, in the latter of which divided 
glands are represented. The crypt layer consists for the most part of the shallower 
crypts for the lodgment of the shorter villi ; but in the section one of the deeper and 
more obliquely directed crypts can be seen. In this deep crypt the bulbous end of 
one of the longer villi, v, is lodged ; p. 645. This and the following figure are from 
drawings by David Hepburn, M.B. 

Fig. 12. Vertical section through the crypt layer, cr, and the more superficial part of the glandular 
layer, gl, of the placental zone ; x 240. The epithelial lining of the shallower crypts is 
seen, and in some of these loose epithelial cells occupy the recesses out of which the 
shorter villi have been drawn. The section has also passed through one of the deeper 
and more obliquely directed crypts, and the villus, v, which it contains has been 
divided. The epithelial covering of the villus and the epithelial lining of the crypt 
are in part in place, but for the most part have been detached from the surfaces which 
they had covered. Closely subjacent to the deeper crypt is a gland, gl, with its 
epithelial lining, which seems as if it might have opened into the crypt ; p. 646. 



Trans. Roy Soc. Edin p -Vol. XXXV. 

W. Turner on Placenta of Dugong. — Plate i. 





M'FarW ^ETskne, Lifll™ Edrn' 






*-* ^ 






Trans Roy. Soc. Edin p -Vo] XXXV. 

W. Turner on Placenta of Dugong. — plate ii 



Fig. 7. 



«*^« ^ ■-■* 




b££ 



-«}- -^m , 



'&f? 






f;$ 



m 



m 





Fig. 5. 




Fig. 6. 









Gustav MaTin. ad. nit. del. 



JKFaplane tErslune. Lith r f Edm r 






Trans. Roy. Soc. Edin r -Vol. XXXV. 

W. Turner on Placenta of Dugong. — Plate hi 



Fiss.12 



MS*. 






Wmm 



',^ ::J ^-k X'i ? 



-t3S* 







9r «■ 






&***&■ 







Fig. 10. 





>y 



Dav.a Hepburn, M.B., ad. nat del. 



M'Fa.rla,tie Jc'Ersline, Litli™ Edin T 



( 663 ) 

XVIII. — Non- Alternate ± Knots, of Orders Eight and Nine. By C. N. Little, of 
Nebraska State University. (With a Plate.) 

(Read 15th July 1889.) 

1. To complete the census of knots of any given order, that is, minimum number of 
crossings, it is necessary to include not only those in which the crossings are taken 
alternately over and under [Alternate ±), but also the Non-Alternate ±, those in which 
two or more consecutive crossings are alike over or alike under. Professor Tait has 
figured the forms of the alternate db knots, of orders three to nine inclusive, on PL XLIV. 
vol. xxxii., Trans. Roy. Soc. JEdin.; the object of this paper is to describe the non- 
alternate knots of these orders. 

2. The projection of a non-alternate knot as a single closed line with double points 
only must be found in the complete series of forms of the alternate knots. The converse 
is not true, and the first operation is to exclude from consideration all forms which are 
not projections of non-alternate knots. 

I find it convenient to use* X and y, as shown in the figure, to give the ^ * 
character of a crossing. The crossing shown looked at as belonging to the d xxxy c 
compartment, or, briefly, part A or B is a X crossing. Looked at as belonging to £^ 
C or D, it is a y crossing. 

3. It is to be remembered that in an alternate knot, or in any portion of *a knot 
where the law of over and under is preserved, the crossings looked at as belonging to the 
parts of either partition (group of compartments) are alike. 

It is evident that in a coil (succession of 2-gons), all the 2-gon crossings are either 
lambda or gamma. 

If two parts are opposite at a crossing, and have besides a connection — as a coil — 
in which all crossings are, say, X, then will the first be a X crossing. It follows that all or 
none of the forms of an alternate will be included among the forms of a non-alternate knot. 

4. It is now easy to decide whether a given alternate form can be the projection of a 
non-alternate knot. For example, in the first form of VI., of the nine folds as shown on 
Professor Tait's plate, the 6-gon amplexum is joined to a 4-gon by a 2-gon, twice by a 
3-gon and by a single crossing, X say. Each 3-gon connection has its three crossings alike, 
and X by § 3 above. If now the single crossing be shifted beyond a 3-gon, the 2-gon 
also is seen to have X crossings, and the law of over and under must hold throughout the 
form. No form then of this knot VI. can be the projection of a non-alternate knot. 

In like manner it is easily seen that the only forms of PL XLIV. , which are the pro- 
jections of non-alternate knots, are the following : — 

Eightfold : I, III, IV, VII, IX, and XIV. 

Ninefold: I, III, IV, V, VII, IX, X, XI, XIII, XIX, XXIII, XXIX, XXXV, 
and XXXVI. 

* Listing called a lambda crossing 8 and used * for a gamma crossing (Vorstudien zur Topologie, p. 52, Gottingen, 1848). 
VOL. XXXV. PART II. (NO. 18.) 5 P 



664 C. N. LITTLE ON NON-ALTERNATE =fc KNOTS. 

5. It now becomes necessary to obtain the complete series of forms of non-alternate 
knots of these orders by assigning to the crossings all possible combinations of X and y. 
In 8III, 8IX, 8XIV, 9IV, 9XIII, 9XIX, 9XXIII, 9XXIV, and 9XXXVI, two parts are 
joined by three connections, in each of which all crossings are alike, and the one, two, or 
three forms can be drawn at once. In each form of 8 III, for example, two parts are 
connected by a 2-gon and twice by a 3-gon. The two forms have y 2-gon, X 3-gon, 
y 3-gon ; and X 2-gon, y 3-gon, y 3-gon. Perversion doubles all numbers where 
amphicheiralism is not found, and will not be again referred to until the final summing up. 
6. Inspection shows that three consecutive overs cannot exist in these 
J orders without degradation of the form ; and this consideration greatly 
shortens the labour of treating the remaining forms. For illustration, 81 
may be taken. Lettering the form as shown, a table is made out where 
+ and — are used for over and under crossings respectively. A portion of the knot in 
which the crossings are alike is enclosed in a parenthesis. 




A 


B 


C 


(D 


E) (F G) A 


H 


C 


(F 


G) B 


H 


(D E) 








1 + 


+ 


— 


— 


+ + - - 


+ 


+ 


— 


+ - 


— 


+ - A 


+ 


B + F + 




2 + 


+ 


— 


— 


+ - + - 


— 


+ 


+ 


— — 


+ 


+ - „ 




F - 




3 + 


— 


+ 


+ 


- + - - 


+ 


— 


— 


+ + 


— 


- + A 


+ 


B - C + D 


+ F + 


4 + 


- 


+ 


+ 


- - + - 


+ 


— 


+ 


- + 


— 


- + „ 


5? 


>) J) » 


F - 


5 + 


- 


+ 


— 


+ H 


+ 


— 


— 


+ + 


— 


+ - „ 




D 


- F + 


6 + 


- 


+ 


- 


+ - + - 


+ 


— 


+ 


- + 


- 


+ - „ 




n D 


- F - 


7 + 


- 


- 


+ 


- + - - 


+ 


+ 


— 


+ + 


— 


- + „ 




„ C - F + 




8 + 


- 


— 


+ 


- - + - 


— 


+ 


+ 


- + 


+ 


- + „ 




F - 





From this table the crossings are marked on the forms. No. 6 is the alternate form. 
No. 3 is an amphicheiral form, but it degrades. No. 7 is the perversion of No. 2, and 
No. 5 of No. 4. This leaves the four distinct forms: 1, 2, 4, and 8. I find 19 eightfold 
forms, and for the ninefolds 88 distinct forms. 

7. Thus far, the work has been straightforward and a matter of routine. In deriving 
the knots from the knot-forms the conditions to be observed are so many that a single 
worker cannot be absolutely certain that all have been observed, and that all the groups 
of forms obtained are really distinct knots. 

I find 3 eightfold non-alternate knots, none of which has an amphicheiral form ; so 
that to the 31 alternate knots already known must be added six new non-alternate, 
making 37 distinct eightfold knots. Professor Tait has shown at N on PL LXXIX. 
vol. xxxii., Trans. Roy. Soc. Edin., five forms of knot I. 

In the ninefolds I find 8 non-alternates, and their 8 perversions, making with the 
82 alternates, 98 ninefold knots. 

These new knots are shown on the Plate. 

8. All of the new knots, with the exception of VIII. of the ninefolds, have single 
degree of beknottedness ; for in every case it is possible to remove in some form of the 
knot at least two crossings by changing the sign of one, and without making the form 
alternate throughout. But there is no non-alternate knot of fewer than eight crossings. 
Knot VIII. has twofold beknottedness. 

H JUL. 90 

(v A 



I. Eight Forms. 





PROFESSOR LITTLE ON N N -A LT E R N AT E KNOTS 
EIGHTFOLD NON-ALTERNATE KNOTS. 












III. Five Forms . 






Trans. Roy. Soc. Ediif, Vol. XXXV 
II. Six Forms. 








I. Twenty two Forms. 

(§(k 



NINEFOLD NON-ALTERNATE KNOTS. 



©f)@)i 



ni(i) 



HI (2) 



III (1) 






&m 



VII (l) VII (2) VII (i) 

II. Two Forms 



XXIV (i 




«& 



III (2) 



h 

vii ( 2 ) ix -- 

II. Sixteen Forms 



IV (1} 




IV (2) 







%> 




XIII (l) XIII (2) 



I 




XIII (3) 



x 111(1) in (z) infi) in (z) iv (1) ivfz) 




rv(3j — v_^ 1V (4) 
IV. Twe nty Form : 



Qg/ to- 

V VII (1) 



III (7) 



.ox $0 ^ 



X!H(l 






i ® i e 

XIII (3) XXIII (l) XXIH(e) 



l©TO®®@@ 



J^ c® 



111(a) 



111(2) 




AA 



X ^-^ XI 

V. Fourteen Forms. 



VII (1) VII (2) 

Q 



VII (,) 



VII (2) IX 



a 



® rfu 



o S3 ^ 



g> ® f @ 



XIII (1) XIII (z) XIII (3) 



XXIII (1) XXIII (z) 



/Vn 




XXIV (1) XXIV (z) 

TV 



2) IV(l) JY^ IV < 3 > IV(t) VII (l) VII(2)' 




XXIII(i) XXIII (2 



XXXV XXXVI 

C H Little del 



in (. 




IV (2) — 1V(3) 

VI. Eight Forms 




ri 



^3 o 



XXIV (1) XXIV (z) 

VII. Four Forms. 



DM 




OCLD 



S, 



X v — ' XI 

VIII. Two Forms. 



(£ © 



K 



XIX(i) 



XIX (2) 



XIX (2) 



/v 





XXXV 



F Hulh.Lith' EJin' 



MM 



The Transactions of the Royal Society of Edinburgh will in future be Sold 

at the following reduced Prices: — 



Vol. 


Price to the 


Price tc 




Vol. 


Price to the 


Price to 


Pnbli 


'rinfc. 


Fellows 


. 


Public 




Fellow. 
£0 16 


>. 

(I 


I. IT. III. 


Out of I 






XXVI. Part 1. 


£1 





IV. 


£0 9 





£0 7 


o 


„ Part 2. 


1 4 





1 





V. 


11 





9 


o 


„ Part 3. 


1G 





12 


6 


VI. 


11 


6 


9 


6 ' 


„ Part 4. 


12 





9 


6 


VII. 


18 





15 





XXVII. Part 1. 


16 





12 





VIII. 


17 





14 





„ Part 2. 


6 





4 


6 


IX. 


1 





17 





„ Part 3. 


1 1 





16 





X. 


19 





16 





„ Part 4. 


1 





16 





XI. 


14 


6 


12 





XXVIII. Part 1. 


1 5 





1 1 





XII. 


14 


6 


12 





„ Part 2. 


1 5 





1 1 





XIII. 


18 





15 





„ Part 3. 


18 





13 


6 


XIV. 


1 5 





1 1 


o 


XXIX. Part 1. 


1 12 





1 6 





XV. 


1 11 





1 6 





„ Part 2. 


16 





12 





XVI. 1 
Part 1. / 


5 





4 





XXX. Part 1 


1 12 





1 6 





„ Part 2. 


16 





12 





Part 2. 


18 





14 





„ Part 3. 


5 





4 





Part 3. 


10 





7 


f) 


„ Part 4. 


7 


6 


5 


8 


Part 4. 


5 





4 





XXXI. 


4 4 





3 3 





Part 5. 


7 





5 


6 


XXXII. Part 1. 


1 





16 





XVII. 


Out of Print. 






„ Part 2. 


18 





13 


6 


XVIII. 


2 2 





1 11 





„ Part 3. 


2 10 





1 17 


6 


XIX. ) 
Part 1. / 


2 2 





1 11 





,, Part. 4. 


5 





4 





XXXIII. Part 1. 


1 1 





16 





Part 2. 


18 





15 





„ Part 2. 


2 2 





1 11 





XX. ") 

Part 1. j 


18 





14 





„ Part 3. 
XXXIV. 


12 





9 


6 


Part 2. 


10 





7 


6 


XXXV. Part 1. 


2 2 





1 11 





Part 3. 


10 





7 


6 


„ Part 2. 


*1 11 





1 3 


6 


Part 4. 


10 





7 


6 


„ Part 3. 


2 2 





1 11 





XXI. | 
Part 1. J 


15 





11 


6 












Part 2. 


10 





7 


6 












Part 3. 


7 





5 


3 












Part 4. 


18 





13 


6 












XXII. 1 
Part 1. J 


1 5 





1 1 















Part 2. 


10 





7 


6 












Part 3. 


1 5 





1 1 















XXIII. ) 
Part 1. / 


15 





11 


6 












Part 2. 


1 15 





• 1 8 


6 


\ 










Part 3. 


1 18 





1 10 















XXIV. ) 
Part 1. j 


1 5 





1 1 















Part 2. 


1 8 





1 3 















Part 3. 


1 10 





1 5 















XXV. ( 


18 





13 


6 












Part 1. J 




















Part 2. 


2 2 





1 11 
















* This Part may be had in Numbers at the following prices : — 

To the Public,— No. 8, 3/-; No. 9, 2/; No. 10, 3/-; No. 11, 2/6; No. 12, 2/6; 
No. 13, 2 ; No. I I, 5 : No. L5, 6/-; No. 16, 1/6-; No. 17, 2/6; No. 18, 1/-. 

To Fellows,— No. 8, 2/3; No. 9, 1 6; No. 10, 2/3; No. 11, 1/11; No. 12, 1/11; 
X<>. 13, 1/6- . No. II, 4/-; No. 15, 1/G-; No. 16, 1 I ; No. 17, 1/11; No. 18, 9d. 



PRINTED DY NEILL AND COMPANY, BDJNBVROH. 




■Hi 
BHHHL 

■Hl 

■■L 

HHHnHta